Regulation of myelopoiesis

INTRODUCTION — The production of myeloid cells (granulocytes and monocytes) is a tightly regulated process. During steady-state hematopoiesis, approximately 10⁸ to 10⁹ of these cells are produced per hour in the bone marrow to maintain their circulating numbers within fairly narrow limits. Production can be rapidly increased in the setting of infection or inflammation. The terms leukocytes and granulocytes are often used interchangeably, although they are somewhat different. The term leukocyte comprises granulocytes, monocytes, and lymphocytes, while the term granulocyte includes polymorphonuclear neutrophils (PMNs), eosinophils, and basophils.

Myelopoiesis begins with the differentiation of a small pool of multipotent hematopoietic stem cells into the most primitive myeloid progenitors. These progenitors develop into recognizable myeloid precursors, which subsequently follow a specific differentiation program that culminates in the emergence of mature neutrophils, eosinophils, basophils, and monocytes (figure 1). This process is most likely driven by expression of successive combinations of transcription factors that in turn dictate the expression of adhesion and hematopoietic growth factor receptors (HGFRs):

●Adhesion receptors play an important role in the localization and release of maturing cells from specific niches in the bone marrow.

●Hematopoietic growth factors (HGFs), such as stem cell factor (SCF, also called Steel factor or c-kit ligand), interleukin (IL)-3, granulocyte-macrophage colony-stimulating factor (GM-CSF), and granulocyte colony-stimulating factor (G-CSF) are important for the amplification of progenitor cells and bind to their target cells through specific receptors.

G-CSF is essential for the amplification and terminal differentiation of neutrophil progenitors and precursors, while macrophage colony-stimulating factor (M-CSF, CSF1) and IL-5 are lineage specific factors for the monocyte and eosinophil lineages, respectively.

This topic will review the elements that underlie granulopoiesis and monopoiesis. A general discussion of hematopoiesis and hematopoietic stem cell function is presented separately. (See "Overview of hematopoietic stem cells".)

ANATOMY OF MYELOPOIESIS

Granulocyte and macrophage colony-forming cells — The first colony assays relevant to the study of the production of granulocytes and monocytes in the mouse were described in 1965 and 1966 [1,2]. Analogous assays were then developed for the human system [3]. These studies demonstrated that individual cells derived from mouse spleen or bone marrow could give rise to colonies of up to several thousand differentiated granulocytes and/or macrophages in a soft agar medium. A period of seven to eight days was required for full maturation of these colonies in the mouse compared with 12 to 14 days in humans.

Appropriate studies were performed to demonstrate the single cell origin of the colonies. These studies also demonstrated that a single progenitor cell, which was termed the colony-forming unit in culture, or CFU-C, was capable of differentiation into both granulocytes and macrophages, thus the designation CFU-GM. Unit gravity sedimentation and other separation methods have been used to demonstrate that CFU-GM represent a cell population distinguishable from the multipotent stem cell [4,5]. A less restricted myeloid stem cell is referred to as CFU-GEMM because it is the precursor to lineage-specific stem cells for granulocytes, erythrocytes, monocytes, and megakaryocytes (figure 1) [6].

Human granulocyte erythroid monocyte macrophage progenitors (CFU-GEMM) can be identified by markers for CD34 and HLA-DR. Some human myeloid progenitor cells express an additional marker for CD64, the Fc-gamma receptor 1, which identifies a specific progenitor cell dedicated to granulocyte and monocyte development (CFU-GM). Micromanipulation of single progenitor cells or two cells separated after the first cell division show multiple combinations of mature myeloid cells can be derived from these progenitors, supporting a stochastic mechanism for the restriction of differentiation potential [7,8]. The major growth factors required for maturing CFU-GEMM into CFU-GM are stem cell factor (SCF), interleukin (IL)-3, IL-6 (induced by IL-3), Flt3 ligand, and granulocyte-macrophage colony-stimulating factor (GM-CSF) (figure 1). GM-CSF receptor expression begins in immature precursor cells and is progressively increased during differentiation into mature granulocytes and macrophages [9].

Long-term liquid bone marrow cultures have been particularly helpful in defining humoral and cell-cell interactions that induce myeloid differentiation [10]. CFU-GM give rise to the more mature granulocyte and macrophage colony-forming units, CFU-G and CFU-M, respectively [1,2]. In addition, CFU-GM can be distinguished from the eosinophil and basophil progenitors, CFU-Eo and CFU-Baso, each arising independently from the common myeloid progenitor cell (figure 1). CFU-GM, under the influence of SF, IL-3, GM-CSF, G-CSF, and/or M-CSF ultimately gives rise to mature phagocytes, the granulocytes and monocytes.

Although a clonogenic assay of granulocyte progenitors was developed almost a decade before the erythroid clonogenic assay, the various factors responsible for granulocyte and monocyte-macrophage development remain incompletely understood.

Neutrophil production and maturation — A model that describes the production and kinetics of human neutrophils suggests that it is highly compartmentalized. The bone marrow contains the most neutrophils (approximately 2.3 x 10⁹ cells per kg) [11]. In contrast, the peripheral blood pool of neutrophils contains approximately 0.7 x 10⁹ cells per kg. The relatively tiny peripheral blood pool is divided into two components in equilibrium: the circulating granulocyte pool and the marginating granulocyte pool. These pools provide entrance into the tissues (figure 2).

The level of circulating cells is buffered by an immense marrow reserve of identifiable precursors. These granulocyte precursors in the bone marrow are also compartmentalized into "mitotic" and "storage" compartments. The mitotic cells are actively dividing (myeloblasts, promyelocytes, and myelocytes), while the storage cells are in a maturing/differentiating, postmitotic state (metamyelocytes, band, and mature neutrophilic forms) (figure 3). The transit times within each compartment are relatively long, so that a huge reserve remains available.

Stages of neutrophil development are characterized by relatively distinct morphological features that are visible in Wright-Giemsa-stained bone marrow aspirates or biopsies viewed by light microscopy. Eosinophils and basophils undergo a similar differentiation program and are first distinguishable at the myelocyte stage when their specific granules are synthesized.

●Myeloblast – Myeloblasts are the earliest myeloid precursors recognizable by light microscopy (picture 1). Their size is 15 to 20 microns; their nuclei are large and round with finely granular chromatin and nucleoli; cytoplasm is scant; and cytoplasmic granules are usually absent.

●Promyelocyte – Promyelocytes are larger than myeloblasts (>20 microns) (picture 2). Their cardinal feature is the presence of many violet-stained (azurophilic; primary) granules with a coarse pattern, often obscuring other structures. The granules are homogenous, dense, and round to ovoid in shape, with a diameter of approximately 0.8 microns. They contain a variety of bactericidal cationic proteins, proteolytic enzymes, and myeloperoxidase [12-14]. Other cellular features (nuclei, nuclear chromatin, nucleoli) are similar to myeloblasts.

●Myelocyte – Myelocytes are the last precursor capable of undergoing cell division (picture 3). They undergo three divisions that result in progressively smaller cells with secondary (specific) granules in addition to the primary granules present at the promyelocyte stage. The specific granules are approximately 0.5 microns in diameter, less dense and more numerous than primary granules, and difficult to see by light microscopy. They contain adhesive proteins such as lactoferrin, integrins, binding protein for cobalamin, and other distinguishing markers as outlined in the table (table 1). Their nuclei become progressively smaller and more irregularly shaped with each division; nuclear chromatin becomes coarse and more clumped, and nucleoli are sparse to absent. The nuclear to cytoplasmic ratio is reduced.

●Metamyelocyte – Metamyelocytes are post-mitotic cells that are slightly smaller than myelocytes and have indented nuclei (picture 4). Their cytoplasm is uniformly pink with finely dispersed granules, similar to that of bands and mature PMNs. Nuclear chromatin is coarse, clumped, and peripherally condensed; nucleoli are absent. Polyribosomes and endoplasmic reticulum (required for protein synthesis and maturation) are also absent.

●Band – Bands (also called stabs; juvenile forms) are smaller cells with elongated, horseshoe-shaped nuclei (picture 5). Their nuclear chromatin is aggregated into evenly arranged clumps. Their cytoplasm is like that of mature PMNs, with pink staining and fine azure/bluish granules. Based on cytochemistry, the ratio of azurophilic primary granules to specific secondary granules is approximately 1:2 [15]. Tertiary (gelatinase) granules are also present. Bands are fully functional phagocytes and may account for 3 to 5 percent of the differential count. They are included in the absolute neutrophil count (ANC) (calculator 1).

●Polymorphonuclear cell (PMN) – PMNs (also called segmented neutrophils) are the fully mature stage of neutrophil development (picture 6). These cells are of uniform size (13 microns in diameter). Nuclei are segmented into two to five lobes (average, three) connected by thin chromatin strands. Nuclear chromatin is coarse and clumped, and stains deep purple. Approximately 3 percent of PMNs from females have a visible Barr body (inactivated X chromosome, visible as a drumstick-shaped mass of dense chromatin 1.5 microns in size, attached to one nuclear lobe by a slender filament) [16]. The cytoplasm contains fine azure/bluish granules and prominent secondary granules, which are finely dispersed and stain the cytoplasm faintly pink. Peroxidase staining reveals peroxidase-positive primary granules and peroxidase-negative secondary granules. Granule heterogeneity is obvious by electron microscopy (picture 7). Mitochondria are scant, and abundant glycogen is present, as the cell depends on anaerobic glycolysis for generating ATP.

Distinction among these stages is often helpful in evaluating and characterizing myeloid malignancies. (See "Evaluation of bone marrow aspirate smears", section on 'High power examination of the marrow'.)

The kinetics of proliferation of recognizable cell precursors have been evaluated using labeled precursors of DNA. These studies indicate that the maturation time for neutrophils from the myeloblast stage is about eight days (figure 3). The so-called labeling indices from which measurements of cell cycle times can be made have served as important approaches to the study of the pharmacology and toxicity of chemotherapy.

Release of neutrophils into the blood compartment — The final stage of granulocyte production, their release from the marrow, is also multifaceted [17]. At least four factors may influence granulocyte egress:

●The organization and localization of the cells in relation to vascular channels

●The development of nuclear and cytoplasmic changes that increase cell deformability

●Factors such as chemokines, cytokines, microbial products, and other inflammatory mediators [18] (see "Approach to the patient with neutrophilia", section on 'Mechanisms' and "Approach to the patient with neutrophilia", section on 'Inflammation')

●The regulation of blood flow through vascular channels in the marrow

The half-life of neutrophils in peripheral blood has been estimated to be 19 hours [19]. Neutrophils then leave the bloodstream and enter tissues to localize at sites of inflammation (figure 2). They interact with the endothelial cells through the integrin receptors, CD11a (LFA-1), CD11b (Mac-1), CD11c (gp150/95), which share a common beta-2 integrin subunit, CD18 [20]. Neutrophils from patients with a total or partial absence of these beta-2 dimeric integrin complexes on the neutrophil surface cannot leave the circulation; patients with leukocyte adhesion deficiency l (LAD l) have a marked neutrophil and monocyte leukocytosis and recurrent pyogenic infections. Other defects in selectin fucosylation and other integrins have been described [21].

Regulation of neutrophil release — The mechanisms that control release of PMNs from the bone marrow are only partially understood. A variety of compounds induce PMN movement into the circulation, including endotoxin, glucocorticoids, a leukocyte-mobilizing factor derived from the third component of complement (C3e), chemoattractants such as C5a, cytokines such as tumor necrosis factor (TNF)-alpha, and certain androgens [22-26]. To emerge into the circulation, PMNs must deform and migrate to the luminal side of the endothelial sinusoidal surface. Endotoxin affects the relationship between marrow sinus endothelial cells and the stromal macrophages covering nonluminal sinusoidal surfaces; it induces retraction of macrophages away from the endothelium which facilitates contact between PMNs and endothelial cells and allows PMN egress from the bone marrow [27,28].

In vitro, PMNs and bands readily pass through filters with pore sizes of 1 micron in response to chemoattractants. Metamyelocytes and late myelocytes are impaired in these responses, whereas early myelocytes, promyelocytes, and myeloblasts cannot move or deform [27,29]. Blood viscosity at high flow rates is a function of cell deformability and volume. As a result, viscosity increases dramatically in patients with markedly increased numbers of circulating leukemic blasts (eg, white blood cell [WBC] count >400,000/microL); this problem does not occur with similar WBC counts due to neutrophils rather than blasts [30].

G-CSF is the principal cytokine that stimulates release of neutrophils from the bone marrow, but the mechanisms that mediate these effects are incompletely defined. SDF-1 (CXCL12), a factor expressed by stromal cells (eg, stromal osteoblasts), and CXCR4, its ligand that is expressed by neutrophils, may play a key role in mediating retention of neutrophils in bone marrow. G-CSF administration leads to a reduction in SDF-1 expression and CXCR4 cleavage [31,32]. In addition, G-CSF increases expression of CXCL2, which is expressed by endothelial cells and binds to its receptor CXCR2 on neutrophils, attracting their migration toward the sinusoidal endothelium [33].

The functional importance of these molecules is illustrated by WHIM syndrome (warts, hypogammaglobulinemia, infections, myelokathexis), in which mutations of CXCR4 cause retention of mature neutrophils in the bone marrow [33], as described separately. (See "Congenital neutropenia", section on 'WHIM syndrome'.)

Neutrophil fate — The turnover rate for PMNs is estimated at 0.3 to 1.3 x 10⁹ cells per kg per day [34]. Studies in healthy volunteers suggest a mean lifespan of approximately five to eight days [35]. The disappearance of labeled PMNs follows a single exponential curve, implying that PMNs are destroyed or leave the blood randomly rather than according to their age (senescence), as is true for erythrocytes and platelets.

The fate of neutrophils includes apoptosis and macrophage engulfment. One candidate for a molecular trigger of apoptosis in PMNs is the product of the FAS (CD95) gene. Fas protein is expressed on the PMN surface, and PMNs constitutively release Fas ligand [36]. Since Fas ligand induces apoptosis in PMNs, its release may provide a paracrine pathway that mediates PMN death [36].

Fas-mediated PMN apoptosis is suppressed in vitro by biologic molecules and drugs (eg, G-CSF, GM-CSF, IL-3, IL-6, IL-15, glucocorticoids) [37-39]. Mediators of inflammation, such as endotoxin, human recombinant C5a, TNF-alpha, and interferon gamma, also inhibit PMN apoptosis in a concentration dependent fashion in vitro [37].

When PMNs become apoptotic in the circulation, CXCR4 is again expressed on their surface and they return to the bone marrow where they are phagocytosed by stromal macrophages. This in turn stimulates G-CSF production, increasing production of PMNs [40].

Under normal conditions, PMNs exit the circulation randomly, followed by their death at inflammatory sites over the following two to three days. Cell death is by a combination of apoptosis and macrophage engulfment [41]. PMNs that transit into the tissues do not return to the circulation.

Monocytes — Monocytes irreversibly leave the circulation [42,43] and differentiate further into fixed-tissue macrophages, a category comprising alveolar macrophages [44], hepatic Kupffer cells [45], dermal Langerhans' cells [46], osteoclasts [47], peritoneal and pleural macrophages, and perhaps brain microglial cells [48].

Dendritic cells (DCs), which bridge the gap between innate and adaptive immune responses, can be derived from granulocyte monocyte progenitors (GMP) at steady state, and include classic DCs (cDC1 and cDC2) and plasmacytoid DCs (pDC). These pDCs have efficient foreign antigen processing capacity and therefore greater T cell stimulating activity. Multipotent progenitors (MPP) give rise to lymphoid-primed multi-potent progenitors (LMPP) that are IL7R+ and also differentiate into pDCs and are less efficient at antigen processing and T cell stimulation than myeloid derived pDCs but have a greater capacity to produce IFN. They also have a 5 to 10 times higher output than the pDCs derived from GMPs. This appears to explain why pDC are more lymphoid in their expression of RAG and rearranged IgH loci, while cDC1s express myeloid antigens that distinguish them from monocytes and cDC2 share the majority of their markers with mo-DC [49-52].

REGULATORS OF MYELOPOIESIS — The enormous diversity of the myeloid system, the high turnover rate of granulocytes, and the necessity to maintain splenic, marginated, and bone marrow granulocyte pools to meet sudden demands caused by infection has led to the evolution of an extremely complex regulatory network. Many of the factors that allow release of granulocytes from storage pools are distinct from colony-stimulating factors (CSFs) [17]. In addition, granulocyte-macrophage CSF (GM-CSF) can inhibit granulocyte motility [53] and can activate granulocytes to increase superoxide anion generation [54].

This complexity in vivo has not permitted studies analogous to those in which erythropoiesis is influenced by hypertransfusion or hemorrhage and megakaryocytopoiesis is stimulated by induced thrombocytopenia; however, in vivo studies using antibodies to neutrophils, germ-free animals, and tissue specific knockouts of regulatory components have begun to provide novel insights into the process by which both emergency and steady-state myelopoiesis are regulated. (See 'Myelopoietic feedback loops' below.)

In early work, investigators relied heavily upon in vitro progenitor assays to study the regulation of myelopoiesis. Both transcription factors and growth factors play an important role in this process and their individual effects have been determined in part by the generation of mice lacking genes for specific proteins (table 2). (See "Regulation of erythropoiesis" and "Biology and physiology of thrombopoietin".)

Transcription factors — Myeloid transcription factors play a major role in regulating the expression of myeloid specific genes and are therefore important for commitment to the myeloid lineage and for normal myeloid differentiation. Whether the transcription factors themselves are regulated by environmental or intrinsic signals is still a major unanswered question. It appears that, once induced, their transcription can often be amplified by autoregulatory loops, and that the factors frequently act in subtle combinatorial mechanisms to induce different myeloid genes [55].

The occupancy of regulatory regions of lineage-specific genes also depends on chromatin configuration and methylation status. There is evidence that during myelopoiesis these epigenetic changes occur in a step-wise manner; in primitive cells lysozyme gene regulatory sites are transiently occupied, whereas in mature macrophages much more stable occupancy is attained at the time that the gene becomes maximally expressed [56,57].

PU.1 — The PU.1 transcription factor is a member of the Ets family and is expressed principally in monocytes/macrophages and B lymphocytes, and also in erythroid cells and granulocytes [58,59]. Potential target genes include the integrin CD11b, M-CSF receptor, GM-CSF receptor alpha, G-CSF receptor, and the Igl light chain [60-64]. Mice that lack PU.1 die in utero with absence of monocytes, granulocytes, and T and B lymphocytes; anemia is variable and therefore does not explain the prenatal mortality [65]. On the other hand, mice with severely reduced expression of Pu.1 (ie, Pu.1 knockdown mice) develop acute myeloid leukemia [66,67] and heterozygous deletion of the Pu.1 locus has been associated with acute myeloid leukemia in humans [68].

C/EBP family — The CCAAT/enhancer binding protein (C/EBP) family of transcription factors bind to DNA through a basic region-leucine zipper domain (bZIP). There are several family members (C/EBP-alpha [CEBPA], C/EBP-beta [CEBPB; also called NF-IL-6], C/EBP-gamma (CEBPG), C/EBP-delta [CEBPD], and C/EBP-epsilon [CEBPE]) that are differentially expressed during myelopoiesis [69]. During maturation of a myeloid precursor cell line to terminally differentiated granulocytes, there was an increase (CEBPB and CEBPE), or an increase followed by either a partial (CEBPD) or marked (CEBPA) reduction in the level of expression of members of this family [70,71]. In one study of 408 patient samples and five cell lines representing 11 different malignancies, mutation of CEBPA was noted in acute myeloid leukemia and myelodysplasia, but rarely in other malignancies [72].

The functions of the different members of C/EBP family have been delineated by target disruption of specific genes:

●CEBPA – CEBPA has been implicated in the regulation of hepatocyte, adipocyte, and myeloid differentiation (levels are low in undifferentiated dividing cells but high in quiescent terminally differentiated cells), and also in myeloid CSF receptor promoter function [73,74]. Cebpa-deficient mice die within the first few hours after birth from impaired glucose metabolism; they also have a selective block in granulocyte production leading to the absence of mature granulocytes and a profound and selective reduction in G-CSF receptor mRNA expression and in CFU-G [75].

●CEBPB – Mice that lack Cebpb produce monocytes that are defective in bactericidal and tumoricidal function; in addition, their macrophages and fibroblasts, but not their endothelial cells, fail to produce G-CSF in response to lipopolysaccharide (LPS) [76]. In another study, Cebpb-deficient mice had an increase in IL-6 production and enhanced proliferation of a number of hematopoietic lineages, a syndrome that resembles Castleman disease [77].

●CEBPE –Mice with null mutations of Cebpe develop normally but die from opportunistic infections at three to five months of age because of functional abnormalities of neutrophils and eosinophils [78]. Cebpe-null neutrophils fail to develop specific granules or their content proteins, and have defects in chemotaxis, respiratory burst generation, and phagocytosis. These observations led to the demonstration that patients with specific granule deficiency, a rare disorder of neutrophil function associated with increased infections, was caused by mutations in the CEBPE gene. The mice also develop a terminal myelodysplasia with proliferation of atypical granulocytes, although it was not clear whether this was truly a clonal disorder or a response to chronic infection.

Core binding factor family — The core binding factor (CBF) family is another group of transcription factors that plays an important role in the regulation of myeloid specific promoters such as those for the myeloperoxidase (MPO) and neutrophil elastase (HNE) genes. These were first identified on the basis of their ability to bind to the polyomavirus enhancer-binding protein 2/core binding factor (PEBP2/CBF). Two alpha subunits (RUNX1; formerly called AML1 or CBF-alpha-1 and CBF-alpha-2), bind DNA with low affinity, but the affinity is strengthened in the presence of the non-DNA-binding beta subunit (CBFB; also called CBF-beta). In addition, RUNX1 can act synergistically with CEBPA to activate the M-CSF receptor promoter [73].

Mice lacking Runx1 die in utero by day 12.5 due to hemorrhage, and they fail to produce all definitive hematopoietic lineages [79,80]. Similar defects have been observed in CBF-beta knockout mice [81]. It has been suggested that RUNX1 and CBFB are required for the emergence of definitive hematopoietic stem cells in the embryo; both genes are also required in the later stages of differentiation of some hematopoietic lineages (megakaryocytes, T and B cells) [82].

These two genes also play a role in leukemogenesis [82,83]. RUNX1 is involved in the most frequent chromosome translocation in acute myeloid leukemia, t(8;21)(q22;q22), which results in the production of the RUNX1::RUNX1T1 (formerly called AML/ETO) fusion protein [84-88], while CBFB is involved in a chromosome inversion inv(16)(p13;q22) that is associated with AML with aberrant eosinophils [89]. The abnormal RUNX1::RUNX1T1 protein can act as a dominant negative inhibitor of RUNX1 [90-92]. However, murine knock-in studies suggest that the mechanism may be more complex. Mice heterozygous for the corresponding gene fusion have a severe block in hematopoiesis, but yolk sac cells generate CFU-M whereas yolk sac cells from Runx1 or Cbfb-knockout mice contain no colony-forming cells [93].

Experiments in mice have demonstrated that Runx1 binds to functionally important sites within the Pu.1 upstream regulatory element and regulates Pu.1 expression at both embryonic and adult stages of development [94]. (See 'PU.1' above.)

Additional evidence for the importance of RUNX1 in myelopoiesis comes from an understanding of Familial AML with mutated CEBPA (familial platelet disorder with predisposition to myelodysplasia and/or acute myeloid leukemia). This is an autosomal dominant disorder characterized by qualitative and quantitative platelet defects, decreased megakaryocyte colony formation, and an approximately 37 percent incidence of AML. Nonsense mutations or intragenic deletion of RUNX1 have been seen in this condition. (See "Familial disorders of acute leukemia and myelodysplastic syndromes", section on 'Familial AML with mutated CEBPA' and "Acute myeloid leukemia: Molecular genetics", section on 'Familial AML'.)

Retinoic acid — Retinoic acid (RA) may play a direct role in myelopoiesis as evidenced by the following observations: vitamin A deficiency is associated with defective hematopoiesis; RA can induce myeloid differentiation in cell lines and in primary leukemic cells; and all-trans retinoic acid is often effective in the treatment of acute promyelocytic leukemia (APL). (See "Clinical manifestations, pathologic features, and diagnosis of acute promyelocytic leukemia in adults".)

The leukemic cells of almost all patients with APL have a balanced (15;17) translocation involving the PML gene on chromosome 15 and RARA (retinoic acid receptor-alpha) gene on chromosome 17. The PML::RARA fusion protein associated with this translocation in APL is an abnormal RAR with altered DNA binding and transcriptional activities, thereby preventing differentiation of promyelocytes. A similar effect, blocked differentiation at the promyelocyte stage, can be induced by expression of a dominant negative RAR in a multipotent murine hematopoietic cell line and in normal mouse bone marrow cells; this effect can be overcome by high concentrations of RA [95]. (See "Molecular biology of acute promyelocytic leukemia".)

Growth factors — A number of growth factors also play a major role in myelopoiesis, including IL-3, GM-CSF, G-CSF, and M-CSF [96].

Effects on progenitor cells — The following observations have been made in murine and human studies:

●Murine IL-3 stimulates a broad spectrum of myeloid progenitor cells including CFU-GM, CFU-G, CFU-M, CFU-Eo, erythroid burst-forming units (BFU-E), megakaryocyte CFU (CFU-Meg) and mast cells.

●As its name implies, GM-CSF was initially thought to provide a more restricted stimulus to the proliferation and development of CFU-GM. However, murine studies with purified or recombinant GM-CSF have shown that it also stimulates the initial proliferation of other progenitors such as BFU-E [97,98].

●The other murine factors, G-CSF and M-CSF are more restricted and predominantly stimulate granulocyte and monocyte colony forming units (CFU-G and CFU-M), respectively [99,100]. However, combinations of stem cell factor (SCF) and G-CSF enhance progenitor cells of multiple lineages in blast colonies, indicating a broader action on primitive progenitor cells [101].

The importance of G-CSF as a myelopoietic growth factor was revealed in mice with deletion of the G-CSF gene [102]. These mice have a chronic neutropenia (20 to 30 percent of normal levels) and reduced bone marrow myeloid precursors and progenitors. They also have a markedly impaired capacity to increase neutrophil and monocyte counts after infection with Listeria monocytogenes.

With the possible exception of GM-CSF, the activities of the human CSFs are similar to those of the corresponding murine factors. Both IL-3 and GM-CSF affect a similar broad spectrum of human progenitor cells including CFU-GEMM, CFU-GM, CFU-G, CFU-M, CFU-Eo, and CFU-Meg (figure 1).

Effects on mature cells — In addition to their effects on progenitor differentiation, the colony-stimulating factors also induce a variety of functional changes in mature cells. As an example, GM-CSF inhibits polymorphonuclear neutrophil migration under agarose [103], induces antibody dependent cytotoxicity (ADCC) for human target cells [53], and increases neutrophil phagocytic activity [54].

Some of the functional changes in mature cells may be related to a GM-CSF-induced increase in cell surface expression of a family of antigens that function as cell adhesion molecules [104]. The increase in antigen expression of CD11b and CD11c is rapid and is associated with increased aggregation of neutrophils; both are maximal at the migration inhibitory concentration of 500 pmol/L, and granulocyte-granulocyte adhesion can be inhibited by an antigen-specific monoclonal antibody. GM-CSF also acts as a potent stimulus of eosinophil ADCC, superoxide production, and phagocytosis [105].

G-CSF and M-CSF also affect mature cells. G-CSF acts as a potent stimulus of neutrophil superoxide production, chemotaxis, ADCC, and phagocytosis [106,107], while M-CSF activates mature macrophages [108] and enhances macrophage cytotoxicity.

GM-CSF — Insight into the in vivo role of GM-CSF comes from studies in mice that carried two null alleles of the GM-CSF gene [109,110]. These mice had normal basal hematopoiesis but developed progressive accumulation of surfactant lipids and proteins in the alveolar space, the defining characteristic of idiopathic human pulmonary alveolar proteinosis. Extensive lymphoid hyperplasia associated with lung airways and blood vessels was also found. (See "Causes, clinical manifestations, and diagnosis of pulmonary alveolar proteinosis in adults".)

Surfactant proteins and lipids are synthesized by type II pneumocytes and cleared from the alveolar space by type II cells and by alveolar macrophages. The lungs from the null animals showed normal surfactant synthetic capacity and no accumulation in type II pneumocytes. In contrast, the alveolar macrophages showed a marked increase in surfactant protein and lipid, suggesting strongly that these cells cannot process surfactant in the absence of GM-CSF.

Similar pulmonary pathology is seen in mice with null mutations of the common chain of the GM-CSF/IL-3/IL-5 receptor (beta c) [102,111]. These mice also have low basal numbers of eosinophils and absence of blood and lung eosinophilia in response to infection with the parasite Nippostrongylus brasiliensis.

Human pulmonary alveolar proteinosis (PAP) is now known to be due to autoantibodies to GM-CSF in the majority of patients [112], while a small number of patients have mutations in the alpha chain of the GM-CSF receptor [113,114] or the common beta chain of the IL-3/GM-CSF/IL-5 receptor [115]. (See "Causes, clinical manifestations, and diagnosis of pulmonary alveolar proteinosis in adults", section on 'Definitions and classification'.)

Innate and adaptive immune responses in rheumatoid arthritis (RA) lead to infiltration of synovial tissue with leukocytes that release TNF, IL-6, GM-CSF, and other cytokines into the tissues. This response attracts other inflammatory cells that mediate tissue damage [116]. GM-CSF levels are increased in the plasma of RA patients and are highly elevated in synovial fluid. These observations have led to investigational studies in RA of monoclonal antibodies directed against the GM-CSF receptor (eg, mavrilimumab, namilumab) [117,118].

Innate and adaptive immune responses in rheumatoid arthritis (RA) lead to infiltration of synovial tissue with leukocytes that release TNF, IL-6, GM-CSF and other cytokines into the tissues. This response attracts other inflammatory cells that mediate tissue damage [116]. GM-CSF levels are increased in the plasma of RA patients and are highly elevated in synovial fluid; these observations have led to exploration of monoclonal antibodies against GM-CSF receptor in this setting [117,118].

Stem cell factor and FLT3 ligand (FLT3LG) — Stem cell factor (SCF) [119,120] and FLT3 ligand (FLT3LG) [121-123], both receptor tyrosine kinase ligands, interact with a variety of hematopoietic progenitor cells, perhaps most importantly with very early stem cell populations (figure 1). SCF also plays an important role in melanocyte growth and development, which is reflected in the coat color effects of mutations in SCF or its receptor, Kit [119].

Mice defective in either Scf (Sl mice) or Kit (W mice) have serious hematopoietic defects including macrocytic anemia, mast cell deficiencies, and deficiencies in the stem cell compartment [119]. In vitro, the proliferative activity of SCF with hematopoietic cells is minimal alone, largely yielding mast cells [124] and is generally most evident when combined with other hematopoietic growth factors (HGFs) [119]. It is particularly effective when combined with HGFs such as IL-3, IL-1 or IL-11 in promoting the expansion of "blast" like cells that retain considerable potential for yielding multilineage colonies in secondary culture [125-127].

The FLT3 receptor tyrosine kinase was originally identified as a novel receptor present in hematopoietic stem cells; in human marrow, its expression is largely limited to the CD34+ cell population [128-130]. FLT3LG alone yields low numbers of CFU-GM colonies from human bone marrow but acts synergistically with other cytokines including IL-3, GM-CSF, erythropoietin, and SCF to yield enhanced colony formation, both in terms of size and numbers of colonies [122,131,132]. The synergy observed between FLT3LG and the other HGFs is comparable to that observed with SCF in similar systems with the exception that FLT3LG has little effect on BFU-E [121].

Homeobox genes — The homeobox genes of the HOXA and HOXB gene clusters are expressed in primitive blood cells and play a role in blood cell differentiation. Disruption of Hoxa9 in mice leads to a 30 to 40 percent reduction in leukocytes and a blunted PMN response to G-CSF [133].

REGULATION OF HEMATOPOIETIC GROWTH FACTOR GENES — Many different cell types, including T lymphocytes, monocytes/macrophages, fibroblasts, epithelial cells, and endothelial cells elaborate various hematopoietic growth factors (HGFs), especially following stimulation [134-137]. In addition, conditioned media from murine organ cultures of lung, muscle, thymus, bone shaft, and heart all contain readily detectable levels of many different HGFs and cytokines [138].

Serum levels of HGFs under steady-state conditions are very low, but elevate rapidly following systemic administration of agents such as lipopolysaccharide (ie, endotoxin) [138]. Thus, it seems likely that normal hematopoiesis is maintained by low level expression of HGFs locally in the microenvironment of the bone marrow, spleen, and thymus, while systemic circulation of HGFs becomes more important following infection when inflammatory mediators such as IL-1 and TNF-alpha can act as potent stimulators of HGF production by many cell types [139-141]. Because HGF production is closely regulated in most cell types, it may be that binding of hematopoietic cells to stromal cells also activates stromal HGF production, analogous to the induction of IL-6 expression by binding of myeloma cells to stromal fibroblasts [142].

Activation of HGF gene expression by various stimuli has been reported for many different cell types [134-137]. This activation serves to accelerate blood cell production in times of stress such as in response to infection, marrow damage, or severe bleeding. Among the many types of HGF-producing cells, the most prominent include T and B lymphocytes, monocytes/macrophages, and mesodermal cells including fibroblasts, endothelial cells, and epithelial cells.

●Expression of some of the HGF genes, notably IL-3 and IL-5, is restricted largely to activated T cells [143-145]. Activated T cells also produce GM-CSF [146], M-CSF [147], and IL-6 [148], but these HGFs are produced by many other cells as well, including monocytes, macrophages, and the various mesodermally derived cells.

●G-CSF is expressed by monocytes/macrophages and the various mesodermally derived cells [137], while the expression of IL-11 is even further restricted, as it does not appear to be made by T cells nor by monocyte/macrophages [149-151].

●Stem cell factor (SCF) and FLT3 ligand are broadly expressed in many mesodermal cell types [119,122,131] but thus far, little is known about the regulation of their expression.

The expression of the HGFs and other cytokines is often triggered in cells by exposure to other cytokines or growth factors, leading to the concept that there are complicated interacting networks of cytokines that serve to control and coordinate many physiologic responses [152,153].

MYELOPOIETIC FEEDBACK LOOPS — Mouse models have demonstrated a granulocyte colony-stimulating factor (G-CSF)-dependent feedback loop that ensures neutrophil production, both in the emergency response to infection and in the steady state (figure 4) [154]. These responses involve G-CSF signals that increase stem and myeloid progenitor proliferation and lead to egress of mature neutrophils from the marrow, the latter achieved by disruption of ligand-receptor interactions through reductions in CXCR4 on the neutrophil surface and CXCL12/SDF-1 in the bone marrow stroma. However, the observations that combined G-CSF and GM-CSF null mice produce neutrophils in the steady state at 20 to 30 percent of normal levels and can increase production in emergencies, such as Candida albicans infection, indicate that non-G-CSF mechanisms are important for both responses [155].

Studies from mouse models have provided new insights into the importance of the G-CSF response in both emergency and steady-state granulopoiesis. Infectious agents are presented to the immune and hematopoietic systems as pathogen-associated molecular patterns (PAMPs) that are recognized by pathogen-recognition receptors, in particular the Toll-like receptors (TLRs). TLRs are expressed by cells of the innate immune system (NK-T cells, gamma-delta T cells, and Th17 T cells) and also by hematopoietic stem and myeloid progenitor cells [156-160]. (See "Toll-like receptors: Roles in disease and therapy".)

Emergency myelopoiesis — The bacterial cell wall component lipopolysaccharide (LPS) is abundant during systemic infection. LPS binds to TLR4, which through a myeloid differentiation adaptor protein Myd88 causes responsive cells to produce G-CSF. Data from Myd88 conditional knockout mice crossed with a variety of tissue-specific Cre animals showed that endothelial cells, but not hematopoietic cells, hepatocytes, bone marrow stromal cells, or pericytes, are essential for the G-CSF response to LPS [161]. Interestingly, the TLRs expressed on hematopoietic stem and progenitor cells, and TLR ligands such as bacterial LPS, can induce self-renewal in HSCs and bias commitment toward myeloid but not erythroid or megakaryocyte lineages [159].

Basal myelopoiesis — Although the control of steady-state myelopoiesis is complex and still incompletely understood, evidence is emerging for several positive and negative regulatory loops [162]:

●Germ-free (GF) mice have neutrophil counts even lower than G-CSF receptor-null animals, and thus LPS or other stimulatory compounds may be supplied by intestinal commensal bacteria (the microbiome) in the steady state to stimulate granulopoiesis through a positive TLR4 mediated G-CSF production pathway [154].

●Further reduction of neutrophils in GF mice by neutrophil antibody treatment leads to another positive feedback process involving TLR4 and an adaptor protein, TRIF, which results in increased G-CSF production [154].

●In a negative feedback loop, tissue macrophages and dendritic cells normally engulf apoptotic neutrophils, which inhibits their secretion of IL-23; this cytokine stimulates innate lymphoid cells to produce IL-17, a cytokine that can stimulate G-CSF expression. Adhesion molecule deficient mice, whose neutrophils cannot transmigrate, show a marked neutrophil leukocytosis and increased levels of G-CSF. The underlying mechanism was shown to involve dis-inhibition of tissue macrophage IL-23 production consequent to decreased neutrophil apoptosis, which thus increases Th17 T cell production of IL-17, leading to G-CSF induction. This led to the hypothesis of a neutrophil "turnstile" mediated by monocyte phagocytosis of transmigrating neutrophils; macrophages thus engaged in apoptosis are postulated to receive anti-inflammatory signals that decrease IL-23 secretion via the nuclear receptors peroxisome proliferator-activator gamma and liver X receptor [163,164].

These mechanisms are probably the tip of an iceberg that will likely involve numerous pathways in the neutrophil response, not only to different bacterial and fungal PAMPs, but also to exposure in different tissues such as the skin, lung, or gastrointestinal tract, and to other non-infectious signals such as inflammation that can result in altered myelopoiesis.

SUMMARY

●Description – Myelopoiesis begins with the differentiation of multipotent hematopoietic stem cells into the most primitive myeloid progenitors. These progenitors develop into myeloid precursors, which follow specific differentiation programs culminating in the emergence of mature neutrophils, eosinophils, basophils, and monocytes (figure 1). (See 'Anatomy of myelopoiesis' above.)

●Differentiation – Maturation of multipotent stem cells is controlled by expression of successive combinations of transcription factors that in turn dictate the expression of adhesion and hematopoietic growth factor receptors. (See 'Regulators of myelopoiesis' above.)

●Cellular contributors – The most prominent cells producing hematopoietic growth factors include T lymphocytes, monocytes/macrophages, and other mesodermal cells, such as fibroblasts and endothelial cells. These factors serve to accelerate blood cell production in times of stress such as in response to infection, marrow damage, or severe bleeding. (See 'Regulation of hematopoietic growth factor genes' above.)

●Growth factors – A variety of compounds induce movement of mature neutrophils into the circulation including endotoxin, glucocorticoids, a leukocyte-mobilizing factor derived from the third component of complement (C3e), chemoattractants such as C5a, cytokines such as tumor necrosis factor (TNF)-alpha, granulocyte colony-stimulating factor (G-CSF), and certain androgens. (See 'Anatomy of myelopoiesis' above.)

●Pattern recognition receptors - Pathogen-associated molecular patterns (PAMPs) that are recognized by pattern recognition receptors, such as Toll-like receptors expressed on both non hematopoietic and hematopoietic cells, dictate both emergency and steady-state myelopoiesis through G-CSF production and through direct effects on hematopoietic stem and myeloid progenitor cells (figure 4). (See 'Myelopoietic feedback loops' above.)

ACKNOWLEDGMENT — UpToDate wishes to acknowledge contributions of the late Laurence A Boxer, MD as a section editor for this topic.

The editors of UpToDate acknowledge the contributions of Stanley L Schrier, MD as Section Editor on this topic, his tenure as the founding Editor-in-Chief for UpToDate in Hematology, and his dedicated and longstanding involvement with the UpToDate program.