DISORDERS OF HEMATOPOIESIS

DISORDERS OF HEMATOPOIESIS

HEMATOPOIESIS

Hematopoiesis is the production of blood cells. It is a tightly regulated system exquisitely responsive to functional demands. The level of neutrophils, eosinophils, and basophils are maintained in discrete ranges with rapid adjustments when demands such as bacterial infection, parasitic infection, or allergic reaction are imposed. Similarly, lymphocytes, monocytes, platelets, and red cells, while maintained in normal ranges, respond rapidly to demands¾lymphocytes to immune challenge, monocytes to various infections, platelets to hemorrhage or inflammation, and red cells to tissue hypoxia from many causes. Derangements in marrow function can lead to an excess of white cells, such as leukemia or leukemoid reactions, or an inadequate number of cells, such as anemia, thrombocytopenia, or leukopenia. Kinetics of cytopenia induction after marrow injury with drugs, radiation, or infections reflect the life span of these cells in peripheral blood. The first lineage to drop are the neutrophils with a blood life span of 6 to 8 h, followed by platelets with a 10-day life span. Anemia develops over a longer time in the absence of blood loss, reflecting the 120-day life span of red blood cells. All of these cell types are produced by primitive cells termed stem cells, which are present in the bone marrow of adult mammals.

The production of all the cell types except lymphocytes is usually very efficient, and production is controlled largely by negative feedback. When demand for production of cells of a particular lineage increases or peripheral levels of the cells fall, stimulatory cytokines are released and generate new cells with a time delay of a few days, or the time required f 555b16f or maturation from stem cell precursors. By contrast, production of lymphocytes is highly inefficient. Each day many more cells are generated than are required in the periphery. Most lymphocytes are destroyed during development; this is due at least in part to the destruction of cells that express antigen receptors specific for self antigens.

HEMATOPOIETIC STEM CELLS

Hematopoietic stem cells are characterized by extensive proliferation and differentiation capacity, with the ability to self-renew on a population basis (Fig. 104-1). They also express a variety of cell-surface proteins and have the ability to rapidly 'home' to bone marrow after intravenous injection. Human stem cells lack markers of lineage commitment (i.e., lineage-negative) and express c-Kit, c-mpl, and usually cluster of differentiation determinant-34 (CD34); a small subset of stem cells may be CD34-negative. Murine cells are also lineage-negative and express c-Kit, CD34, c-mpl, and Ly6A or Sca. The most primitive cells are characterized by low-level expression of a relatively large number of cytokine receptors and by relative exclusion (or pumping out) of the dyes rhodamine and Hoechst. These cells express a variety of adhesion proteins presumptively involved in marrow homing, including alpha4, alpha5, alpha6, L-selectin, and platelet/endothelial cell adhesion molecule (PECAM) (Fig. 104-2). Another characteristic of the stem cell is a functional plasticity in response to cytokines as it transits the cell cycle (Fig. 104-3). Engraftment capacity is good in G1 but virtually lost in late S and early G2.

The stem cell is also a highly mobile cell with the capacity to evolve rapidly or involute pseudopodial extensions. The gold standard for defining the stem cell has been in vivo repopulation and long-term reconstitution of lethally irradiated mice. In vivo repopulation studies using unique radiation-induced chromosomal abnormalities or retroviral markers have shown that one or, at most, a few stem cells are capable of reconstituting the entire lymphohematopoietic system of a mouse; they also have defined classes of stem cells with short, medium, or long-term repopulating capacity. These are cell types that differ in the kinetics of hematopoietic reconstitution. Short-term cells repopulate in the first few weeks after transplantation but are not long-lasting; long-term cells account for long-lived reconstitution beginning a few months after reconstitution and lasting the entire life span; medium-term cells bridge the time between short- and long-term cells. When relatively small numbers of marked stem cells¾obtained by limiting dilution of sorted marrow cells¾are transplanted, lymphohematopoiesis may be clonal or oligoclonal, initially. Normal polyclonal lymphohematopoiesis derives from a relatively large number of clones. Both competitive marrow repopulation and mathematical studies support the model of polyclonal hematopoiesis. The most primitive long-term repopulating cells on activation with cytokines can rapidly alter phenotype and become short-term repopulating cells.

While stable multilineage chimerism has been documented in humans after clinical marrow transplantation, no assay system exists for human stem cell activity. A number of surrogate assays are used for the long-term, multilineage-repopulating cell in both humans and mice. These include the multifactor-responsive, high-proliferative potential colony-forming cells (HPP-CFC) and variations of stromal-based assays including the cobblestone-forming cell, long-term culture-initiating cell (LTC-IC) or LTC-IC-extended (LTC-IC-e). The adequacy of these assays is still the subject of debate. In addition, the NOD-SCID immunodeficient mouse has become a surrogate model for assaying human hematopoietic stem cells, although lineage skewing and variability of engraftment undermine its reliability.

LINEAGE PLASTICITY OF STEM CELLS

Tissue stem cells are capable of producing a wide variety of differentiated cell lineages, depending on intrinsic cell programming and the microenvironmental signals. Marrow cells may differentiate into mesenchymal, myocyte, endothelial, hematopoietic, and neural cells. Neural muscle and hepatic stem cells have been reported to give rise to hematopoiesis in transplanted mice. The regulation of stem cell plasticity and life span remains incompletely understood. However, once a particular set of transcription factors has been induced, either through an intrinsic program or from extracellular signals, reversibility is limited. The sequentially ordered activation of transcription factors leads to lineage commitment.

MICROENVIRONMENT

Nonhematopoietic tissues exert major influences on hematopoiesis, both short- and long-range. The nonhematopoietic tissues immediately abutting hematopoietic tissue have been termed the hematopoietic microenvironment, and the cells that comprise the environment influence hematopoiesis. For example, a surface location of adoptively transferred murine stem cells in the spleen of lethally irradiated mice favored erythropoiesis, while an intrasplenic trabecular location was biased toward granulocyte production. In both human and murine species, various cell types have been identified in stroma, including hematopoietically derived macrophages and nonhematopoietic preadipocytic fibroblasts, endothelial cells, and vascular smooth muscle. This system appears capable of supporting the most primitive stem cells and controlling their proliferation and self-renewal. Most stem cells are resting under normal physiologic conditions but can be recruited into the cell cycle by demands of increased terminally differentiated hematopoietic cells. Cell-cell contact is critical in determining the microenvironment stimulus. Stem cells and primitive cells bind tightly to the stroma, while maturing precursors and terminally differentiated cells are nonadherent. Blocking interactions between stem cells and stromal cells with antibodies to vascular cell adhesion molecule (VCAM)-1 on stromal cells or its ligand, VLA-4, on stem cells block the interaction. Cytokine receptors binding to membrane-associated cytokines like stem cell factor, or to extracellular matrix-bound ligands, contribute other adhesive interactions.

PROGENITORS

Bone marrow stem cells can be induced to proliferate and differentiate into a wide variety of mature cell types in vitro in the presence of an appropriate colony-stimulating factor (CSF). Cells that give rise to mature colonies of granulocytes and macrophages are called granulocyte-macrophage colony-forming units (CFU-GM) (Fig. 104-4). The particular hematopoietic growth factor that stimulates the development of these colonies is called granulocyte-macrophage colony-stimulating factor (GM-CSF). Distinct culture conditions and supplemental growth factors, alone and in combination, are capable of producing a range of cell expansions from multilineage colonies that include lymphocytes to single-lineage clones. The different stem/progenitor clones are summarized in Table 104-1. Progenitor cells in general are found to have a higher proliferative rate and more lineage restriction than stem cells. They are also responsive to smaller numbers of cytokines. Thus, they are defined by expression of a limited variety of cytokine receptors.

The size of the colonies denotes the activity of cells at different stages of differentiation. Terminally acting cytokines produce smaller colonies called CFU (colony-forming units). When progenitors are stimulated with mixtures of early- and late-acting cytokines and are cultured for longer periods of time, the colonies are larger and multiple lineages are represented. Primitive multifactor-responsive erythroid colonies are termed burst-forming unit erythroid (BFU-E) while even more primitive colonies with great proliferative potential are termed HPP-CFC.

CYTOKINES

The lymphohematopoietic stem/progenitor populations and their progeny are largely defined by their cytokine responsiveness and cytokine receptor phenotype. Major efforts to define the regulators of granulocyte, erythroid, and platelet production have culminated in the definitions of a variety of glycoproteins. Acting through cell surface receptors at very low concentrations, these glycoproteins control the production of stem cells in vivo. Most prominent have been erythropoietin for red blood cells, GM-CSF for granulocytes and macrophages, granulocyte-CSF (G-CSF) for granulocytes, and thrombopoietin for platelets. In addition, macrophage-CSF or CSF-1, was defined as a primary regulator of macrophage-monocyte production and function. These cytokines exert prominent actions on specific cell lineages, but all exert actions on different cell lineages or on cells that have the potential to differentiate along more than one lineage.

In addition to the more lineage-restricted cytokines, a large number (perhaps up to 70) act broadly on multiple lineages and at multiple stages of lymphohematopoiesis. They exert effects on renewal, proliferation, survival, and differentiation; these effects may be stimulatory or inhibitory, and the cytokines usually show additive or synergistic effects with other cytokines. The cytokines also modulate intrinsic functions of early stem cells (migration and cell adherence) and promote the effector functions of their terminally differentiated progeny. G-CSF primes neutrophils to undergo oxidative metabolism in response to formyl-methionyl-leucyl-phenylalanine (fMLF) and enhances cell migration, while interleukin (IL) 3 activates basophils, mast cells, and eosinophils. CSF-1 at low levels supports survival of murine marrow macrophages and at higher levels stimulates protein synthesis, cell division, and various macrophage functions, including antitumor activity, secretion of products of oxygen reduction, and plasminogen activator. CSF-1 also induces secretion of IL-1 from macrophages. Many of the hematopoietically active cytokines induce secretion of other cytokines, either inhibitory or stimulatory, creating multiple cytokine regulatory loops. Transforming growth factor b (TGF-b) is an inhibitory cytokine but also an autocrine factor supporting survival of pluripotent hematopoietic stem cells by blocking G1 to S phase transition. TGF-b conversely shows stimulatory effects on progenitors. Cytokines also modulate adhesion protein and integrin expression on multiple cell types. They exert their effects by interacting with surface-based receptors and initiating second-messenger cascades (see below).

The lymphohematopoietic cytokines can be broadly divided into colony-stimulating factors, erythropoietin, thrombopoietin, the interleukins, the inhibitory cytokines, chemokines that regulate cell migration and activation, and a variety of other hematopoietically active cytokines. A noninclusive overview of these cytokines emphasizing their primary, highlighted, or first-described action is presented in Tables 104-2, 104-3 and 104-4. The general characteristics of cytokines are summarized in Table 104-5.

CYTOKINE RECEPTORS, SIGNAL TRANSDUCTION, AND TRANSCRIPTION FACTORS

Cytokines induce their effects through cell-surface membrane receptors. Several cytokine receptor families have been identified. The hematopoietic receptor family includes IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, G-CSF, GM-CSF, and erythropoietin. Common characteristics of this family include four conserved cysteine residues and a WSXWS motif (X is a variable, nonconserved amino acid). Some also have immunoglobulin-like structures in their extracellular domains. Receptors frequently consist of multiple chains, and dimerization on cytokine binding is a usual feature of receptor biology. These receptors have no intrinsic signaling capacity and transmit signals by attaching to intracellular signaling molecules, such as the src family and the JAK family kinases. GM-CSF, IL-3, and IL-5 receptors have low-affinity alpha chains and a common high-affinity beta chain. The common beta chain may play a role in the competitive binding of these ligands.

Receptors for FLT-3 ligand, c-kit, platelet-derived growth factor (PDGF), CSF-1, and thrombopoietin constitute the tyrosine kinase receptor family. These receptors have conserved cysteines in the extracellular domain, with tyrosine kinase activity in the cytoplasmic domain, an immunoglobulin-like structure involved in ligand, and binding. Chemokine receptors are seven-transmembrane (serpin) G-protein linked receptors that signal cell activation and migration.

Cytokines typically cause receptor oligomerization on hematopoietic cells, followed by activation of intrinsic (receptor) or extrinsic tyrosine kinases, phosphorylation of the receptor and recruitment of Src-homology (SH2), and phospho-tyrosine binding (PT3) domain proteins to the receptor. Subsequent steps vary with different cytokines but essentially represent a series of phosphorylation-dephosphorylation events, with the final activation or nuclear translocation of a protein or protein complex that binds specific regions of DNA and initiates various genetic programs (i.e., acts as a transcription factor).

The complexity of these second-messenger signaling systems is illustrated by signaling through the GM-CSF, IL-3, and IL-5 receptors, which share a common beta chain. The beta chain does not have kinase activity but induces tyrosine phosphorylation of itself and a number of cytoplasmic proteins, including kinases, such as P1-3 kinase; adapters illustrated by Grb2; the insulin receptor-substrate 2 Cbl and Shc; guanine nucleotide exchange factors such as Vav; phosphastases such as SH2-domain protein tyrosine phosphatase-2 and SH2-containing inositol phosphatase; and transcription factors such as STAT 5.

Receptor phosphorylation is mediated by receptor-associated kinases, such as JAK2 (Janus family kinase 2, named Janus for the Roman god who guards the gates and looks in two directions; original Janus kinases were felt to have both tyrosine and serine kinase activity) and Src-family kinases. These sequential protein interactions lead to the evolution of proteins or protein complexes, termed transcription factors, that bind to specific regions of DNA to initiate genetic programs determining survival, proliferation, differentiation, and function.

As with second messengers, the transcription factor field is complex and evolving, but a number of transcription factors associated with specific stem cell levels or differentiation pathways have been described. Transcription factors that act at the earliest stem cell levels include c-myb, p45-NF-E2, GATA-2, AML-1 and tal-1/SCL, while Ikaros and PU-1 may act at the earliest lymphoid level. GATA-1 influences erythroid, mast cell, and megakaryocyte lineages, while FOG (friend of GATA-1) acts in concert with GATA-1. PU-1 appears to influence granulocyte and monocyte differentiation, P45-NF-E2 affects megakaryocyte lineages, and PAX-5 B lymphoid development. These transcription factors usually act in complexes with specific conformations binding to particular DNA sequences.

MIGRATION HOMING AND ADHESION PROTEINS

The process of stem cell homing to the marrow is complex and involves a number of adhesion proteins. Very late antigen (VLA) 4, VLA-5, VLA-6, PECAM, P- and E-selectin, CD44, CXCR, and a receptor for ligand-bearing galactosyl and mannosyl residues have been shown to be expressed by hematopoietic stem/progenitor cells and implicated in marrow homing. The integrins a4 and/or a5 are expressed on immature blasts, erythroid progenitors, monocytes, and CD34+ cells; in general, expression of a4 appears to decrease with maturation. Hematopoietic cells also bind differentially to different extracellular matrix components: erythroid cells to fibronectin, CFU-GM and BFU-E to collagen.

Antibody to VLA-4 given in vivo causes mobilization of hematopoietic progenitors in normal or cytokine-treated primates and/or mice. Stem cell mobilization by cytokines involves down regulation of adhesion protein expression on hematopoietic stem cells. The cell-cycle related fluctuations in engraftment appear to be based on alterations on different surface adhesion proteins. The stem cells are highly motile and move in a direction of cytokine or chemokine gradients with steel factor and stromal-derived factor 1 (SDF-1) being active. Adhesion proteins act not only for motility/adhesion, but also serve a regulatory role that is similar in some cases to traditional cytokines.

PHYSIOLOGY OF HEMATOPOIESIS AND SOURCES OF CYTOKINES

Erythropoietin is produced largely by the kidney in response to tissue hypoxia. The regulation of granulocyte and monocyte production is more complex, but appears to be in response to various infectious or noxious agents, such as gram-negative bacteria, the endotoxin in the cell wall of these bacteria, and antigen stimulation. All of these interact with peripheral tissue cells to generate a variety of cytokine messages, resulting in increase production in specific cell types. Parasitic infections appear to elicit IL-5, which modulates the eosinophilia and mast cell lineages. Viral infections have specific effects on lymphocyte classes; typically bacterial infections stimulate granulocyte production. Tuberculosis or other mycobacterial infections predominantly induce increased monocyte production. All of these biologic affects appear to be mediated by the selective evolution of cytokine complexes from tissue endothelial cells, fibroblasts, macrophages, and lymphocytes. Most cells produce a large variety of cytokines, but the key is the relative levels, combinations, and timing of the production of these cytokines (Fig. 104-5).

HEMATOPOIETIC STEM CELL AND CYTOKINE DISEASES

The classic stem cell disease is chronic myeloid leukemia. Here a specific genetic translocation between chromosomes 9 and 22 at the stem cell level leads to excess production of granulocytes, monocytes, basophils, frequently platelets, and less frequently red cells. Other lymphohematopoietic clonal stem cell diseases include polycythemia vera, myelofibrosis with myeloid metaplasia, paroxysmal nocturnal hemoglobinuria, and acute myeloid leukemia. Out of the scope of this chapter, but relevant to these discussions, is the fact that many lymphoid neoplasms are clonal diseases at early stages of development, but probably not at the mature stage suggested by the tumor cell-surface phenotype. The vast majority of peripheral B cell and T cell malignancies have genetic lesions associated with receptor gene rearrangements, which occur early in lymphoid cell development. Aplastic anemia appears to be a disease characterized by a defective number of hematopoietic stem cells. Cyclic hematopoiesis is another disease of hematopoietic stem cells. In gray collie dogs with this disorder, levels of platelets, reticulocytes, monocytes, and granulocytes cycle. This disease can be cured or transmitted by marrow transplantation. The human disease, cyclical neutropenia¾or cyclic hematopoiesis in which blood cells oscillate with a 21-day period¾is caused by missense and splicing mutations in the gene encoding neutrophil elastase, thus implicating this inflammatory chymotryptic serine protease in the oscillatory timing of hematopoiesis. The stem cell diseases are summarized in Table 104-6.

A number of cytokine disorders or diseases have now been defined. The best characterized is the anemia of renal failure, an erythropoietin deficiency state that can be corrected by the administration of erythropoietin. Various tumors, particularly lung cancer, increase peripheral granulocyte counts secondary to the production of G-CSF. The IL-6 family of cytokines appears to be prominently involved in a number of inflammatory states, causes the systemic symptoms associated with Castleman's disease and atrial myxoma, and may be an etiologic factor in multiple myeloma. IL-6 may also be a major cause of symptoms in various lymphomas. Abnormalities of the c-Kit receptor may underlie a number of mast cell diseases in humans; IL-5 production is the proximate cause of a number of eosinophilic states. A deficiency of IL-1 is a feature of aplastic anemia. Mutations in the G-CSF receptor in chronic congenital neutropenia (Kostmann's syndrome) may be a causative factor in the evolution of acute myeloid leukemia in some of these patients.

THERAPEUTIC IMPLICATIONS OF STEM CELLS AND CYTOKINES

Stem Cells Stem cell transplantation was first established as an effective therapy for relapsed acute myeloid leukemia and aplastic anemia. It is now a mainstay of therapy for virtually all leukemias and some relapsed lymphomas. Application of this treatment to a number of solid tumors has been disappointing. Major expectations with regard to its potential in breast cancer have not yet been fulfilled, although it appears to be effective in relapsed testicular cancer. The rationale is the use of very high doses of drugs or radiation designed to kill all tumor cells, but at levels where marrow toxicity would be lethal. Marrow damage is the dose-limiting toxicity for many chemotherapeutic agents. If marrow function is replaced by transplant, it might be possible to increase the dose of chemotherapy substantially before another organ toxicity becomes dose-limiting.

The strategy is somewhat different in intrinsic marrow diseases, such as aplastic anemia, where marrow function is restored without the need for killing tumor cells, or in genetic marrow diseases such as thalassemia and sickle cell anemia, where replacement of the abnormal marrow with normal marrow corrects the disease state. Aggressive autoimmune diseases are also being treated with marrow replacement. Stem cells can come from a related or unrelated allogeneic source or directly from the patient. The sources of the stem cells also vary. Initially, marrow aspirate was the predominant source, but now apheresed peripheral blood stem cells are the most utilized. In addition, umbilical vein cord blood, especially in pediatric patients, appears effective. Numbers of cells are sometimes limiting for adult recipients (Chap 115).

Cytokines Demonstration that hematopoietic cytokines could modulate red blood cell and white cell production in humans has been useful in some clinical settings. Erythropoietin treatment improves hematocrit and quality of life in patients with chronic renal failure. Erythropoietin has been tried in myelodysplastic syndromes (MDS), a group of clonal stem cell disorders. Meta-analysis shows an overall response rate of only 13%; the actual overall clinical benefit was exceedingly small. In addition, the best results were seen in patients receiving daily injections. Prohibitive costs and often poor response limits use to those patients with serum EPO levels lower than 500 mU/L, and less than 5% myeloblasts. The utilization of erythropoietin in other settings than renal failure remains controversial and its use may relate more to effective marketing than to the science.

These concerns are multiplied for the use of the myeloid growth factors, G-CSF, and GM-CSF. These agents elevate neutrophil and monocyte counts, and under very selective conditions they can result in a reduced toxicity of various chemotherapeutic regimens (Table 104-7). Unfortunately, they save about the same amount of money in hospitalizations for febrile neutropenia as they cost, and their use has not increased survival rate. Virtually all of the G-CSF and GM-CSF trials in cancer patients have been flawed by design; they involve escalation of drugs to toxic levels and reversal of toxicity without addressing the question of whether the patient's survival is affected by the treatment. GM-CSF and G-CSF are grossly overutilized; they are often used in settings where their efficacy has not been shown (e.g., patients with a low probability of neutropenia). Their use should still be considered experimental, and they should continue to be studied in a protocol setting but not used routinely. G-CSF is useful in mobilization of stem cells, and it is also effective in treatment of various chronic neutropenias, in particular cyclic neutropenia and Kostmann's syndrome. G-CSF may be involved in the evolution to acute myeloid leukemia in some patients, but overall it appears to be an effective intervention in these seriously ill patients. G-CSF may also aid in healing of diabetic skin ulcers.

IL-11 and thrombopoietin can elevate platelet counts in experimental animals, but their place in clinical practice is unclear. IL-11 has been approved for use in chemotherapy-induced thrombocytopenia, but its effects are small. Use of pegylated recombinant human megakaryocyte growth and development factor¾the truncated version of thrombopoietin¾has resulted in production of neutralizing antithrombopoietin antibodies and thrombocytopenia. Recombinant human thrombopoietin does not commonly elicit neutralizing antithrombopoietin antibodies. Clinical benefit (or cost effectiveness) has not yet been shown with thrombopoietin. Surrogate values of platelet counts or number of platelet transfusions are not valid criteria for clinical benefit. Thrombopoietin may eventually find a role as an expander of early stem cells in vitro. Active research in this important area continues.

Gene Therapy Hematopoietic stem cells provide an ideal vehicle for various gene therapy approaches. These cells can be easily induced into the cell cycle for retroviral integration. Long-term expression of introduced genes is currently being obtained in animal models. Initial clinical application has been disappointing, but success has been obtained in Gaucher's disease, suggesting that gene therapy will eventually become a successful approach to a number of hematopoietic diseases.