Section 5. Inflammatory Diseases Based on Neutrophils and Eosinophils

This section presents an overview of neutrophil biology and function and uses a few well-characterized defects of myeloid function as illustrations.

Ontogeny and Development

Similar to other components of the hematopoietic system, the neutrophil is ultimately derived from a pluripotent hematopoietic stem cell. The development of the myeloid stem cell is largely determined by ambient cytokines and reflected in its surface markers, morphology, and functional characteristics. The myeloblast is fully committed to the neutrophil lineage and is the first morphologically distinct cell in neutrophil development. Subsequent stages of neutrophil development occur under the influence of granulocyte colony-stimulating factor (G-CSF) and granulocyte–macrophage colony-stimulating factor (GM-CSF). Four to six days are required for maturation through the mitotic phase to the myelocyte, and 5–7 days more for the myelocyte to develop into a mature neutrophil, including the metamyelocyte and band stages, before emerging as a fully developed neutrophil. Development of neutrophils through the myelocyte stage normally occurs exclusively in the bone marrow, which is composed of approximately 60% developing neutrophils. The mature neutrophil measures 10–12 μm and has a highly condensed, segmented, multilobulated nucleus, usually with three to five lobes. Although 1011 neutrophils are generated daily, this number can rise tenfold in the setting of infection. The calculated circulating granulocyte pool is 0.3 × 109 cells/kg blood and the marginated pool is 0.4 × 109 cells/kg blood, comprising only 3% and 4% of the total granulocyte pool, respectively. The bone marrow releases 1.5 × 109 cells/kg blood/day to this pool but keeps 8.8 × 109 cells/kg blood in the marrow in reserve. An additional reserve of immature and less competent neutrophils, 2.8 × 109 cells/kg blood, is also available.

G-CSF is critically important for neutrophil production.1 Mice deficient in G-CSF show reduced neutrophil numbers and cannot upregulate neutrophil numbers in response to infection. Interestingly, G-CSF production is under the influence of IL-17, a cytokine of importance in regulation of epithelial defenses.

Biologic Functions

Granule Content and Function

(Table 30-1.) Neutrophils are characterized by cytoplasmic granules and partially condensed nuclei. Granules are first found at the promyelocyte stage.2 Primary (azurophilic) granules are the first to arise, measure approximately 0.8 μm in diameter, and contain numerous antimicrobial products including lysozyme, myeloperoxidase, and defensins.3 Primary granules are only synthesized at the promyelocyte stage. The promyelocyte gives rise to the myelocyte, the last cell of the neutrophil lineage with proliferative potential. Therefore, cytokines or agents that increase total neutrophil production must act at or before the myelocyte stage. The smaller eosinophilic secondary (specific) granules appear during the myelocyte stage. These granules measure about 0.5 μm in diameter and contain lactoferrin, collagenase, gelatinase, vitamin B12-binding protein, and complement receptor 3 (CR3; CD11b/CD18). gp91phox and p22phox comprise the specific granule component cytochrome b558, defects in which cause chronic granulomatous disease (CGD), characterized by infections with particular catalase-producing bacteria. Gelatinase also cleaves and potentiates the activity of the chemokine interleukin-8 (IL-8). Because primary granules are synthesized early and distributed to daughter cells during division, they are eventually outnumbered by about 3:1 by the specific granules, which are produced throughout the myelocyte stage.

Granules fuse in a sequential fashion with incoming phagocytic vacuoles, such as those containing ingested bacteria. Secondary granules fuse to the phagosome within the first 30 seconds after ingestion and release their enzymes, many of which function best at neutral or alkaline pH. By 3 minutes after ingestion, the primary granules have fused to the phagolysosome leading to rapid lowering of the intravacuolar pH. For objects too large to be ingested, or certain stimuli, degranulation to the cell surface occurs with release of granule contents into the surrounding environment. This can be inferred by detection of lactoferrin levels in blood.

An example of disordered granule biogenesis is Chédiak–Higashi syndrome (CHS), a rare autosomal recessive disorder with abnormal pigmentation due to a generalized abnormality of primary granule and lysosome formation (see Chapter 143).

Neutrophil-Specific Granule Deficiency

Neutrophil-specific granule deficiency is a rare, autosomal recessive condition clinically characterized by a profound susceptibility to bacterial infections. There is a paucity or absence of neutrophil-specific granules, specific granule proteins (e.g., lactoferrin) and their respective messenger RNAs, and very low levels of the primary granule products defensins and their messenger RNAs. Specific granule deficiency is due to loss of the transcriptional factor CCAAT/enhancer binding protein ε (CEBPε), which is essential in normal myeloid development. Acquired abnormalities of neutrophil granules are seen in some myeloid leukemias, in which primary granule contents may be aberrantly accumulated (e.g., Auer rods in acute myelogenous leukemia).

Tissue Trafficking

Chemoattractants and Chemotaxis

Metchnikoff discovered over a century ago that neutrophils move toward very slight gradients of chemical signals, now termed chemoattraction. The “classic” chemoattractants are N-formylmethionyl-leucyl-phenylalanine (fMLF), complement factor 5a (C5a), leukotriene B4, and platelet-activating factor (PAF). More recently, chemokines (chemoattractant cytokines), a class of small (<10 kDa) extremely active chemoattractant proteins have been identified (see Chapter 12).

IL-8 is a potent chemoattractant and neutrophil activator (see Chapter 12). Its blockade by neutralizing antibody in rabbit models of pulmonary inflammation prevented neutrophil accumulation and reperfusion injury, confirming its central role in neutrophil recruitment to sites of inflammation. Cellular sources for IL-8 include neutrophils, monocytes, T cells, B cells, natural killer cells, basophils, eosinophils, fibroblasts, endothelial cells, keratinocytes, and smooth muscle. Exudative neutrophils are particularly effective at synthesizing IL-8, probably through a calcium-regulated mechanism.

Exuberant neutrophil function is implicated in Sweet syndrome (acute febrile neutrophilic dermatosis) as shown by the frequent association of the syndrome with hematologic malignancies, dysmyelopoiesis, and sometimes with the use of G-CSF (see Chapter 32).

The classic chemoattractants and chemokines use similar receptors that have in common seven transmembrane regions, an extracellular amino terminal domain, and transduce their signals through pertussis toxin-sensitive heterotrimeric G proteins. These are the same type of receptors through which light, neuropeptides, and neurotransmitters signal. Chemoattractant receptor signaling entails exchange of bound guanine diphosphate for guanine triphosphate by the α subunit of the heterotrimeric G protein, which in turn leads to the dissociation of the β–γ subunit of the complex, stimulation of phospholipase C, and generation of inositol triphosphate and diacylglycerol from phosphatidylinositol bisphosphate. Inositol triphosphate stimulates the release of calcium from intracellular “calciosomes,” whereas diacylglycerol activates protein kinase C. Extracellular calcium also enters the cell, preparing it for subsequent movement, generation of oxidants, and secretion of vesicles.

Ras-related guanosine triphosphatases of the ρ superfamily are also involved in actin cytoskeletal regulation and adhesion. Phosphatidylinositol 3-kinase is activated by rho and ras and is necessary for the neutrophilic respiratory burst, adhesion, endothelial transmigration, and chemotaxis. Mitogen-activated protein (MAP) kinases are serine/threonine kinases that include p38, Erk1, Erk2, and Jnk and are involved in neutrophil signaling and adhesion. Inhibition of p38 impairs chemotaxis and tumor necrosis factor (TNF)-α-mediated superoxide production, adhesion, and release of secondary granules. Erk activation is required for neutrophil homotypic aggregation. Salicylates inhibit neutrophil adhesion through Erk inhibition.

Several chemokine receptors also function as requisite human immunodeficiency virus (HIV) coreceptors. Mutations in the chemokine receptor CCR5 that are protective against HIV infection led to development of chemokine receptor blockers for treatment of HIV infection (maraviroc and vicriviroc).

Adhesion

Neutrophils exist as free-flowing (those which are sampled on blood drawing) and marginated cells (those which are attached to the endothelium or are traversing the lung, skin, or other tissues). Neutrophils rolling along the endothelium recognize sites of activation (e.g., chemokine expression), adhere to those sites, and traverse the endothelium to enter the tissue and fight infection. Leukocyte physical interaction with endothelium and other leukocytes is mediated by integrins, selectins, and intercellular adhesion molecules (ICAMs; Fig. 30-1).

Elaboration of chemoattractants or display of activation markers on endothelium triggers leukocyte high affinity binding by β2 integrins, heterodimeric surface molecules largely stored in the secondary granules of neutrophils that are displayed on the cell surface upon leukocyte activation. There are three β2 integrin heterodimers comprised of different α chains, CD11a, -b, and -c, and a common β chain, CD18. Each CD11/CD18 complex has separate and overlapping activities. CD11a/CD18 [leukocyte function-associated molecule 1 (LFA-1)] binds to other leukocytes and mediates tight adhesion to the endothelium through ICAM-1 and ICAM-2. CD11b/CD18 (Mac-1, Mo-1, or CR3) binds to the inactivated form of the third component of complement (C3bi) and thereby facilitates complement-mediated phagocytosis. CD11b/CD18 also binds to bacteria directly, to fibrinogen, and to endothelium through ICAM-1. The divalent cations Ca2+ and Mg2+/Mn2+ mediate adhesion through β2 integrin “A” domains containing a metal ion-dependent adhesion site. CD11b/CD18 may also induce the expression of the β1 integrin very late antigen 6 [VLA-6 (CD49f/CD29)], derived from neutrophilic granules, to aid in tissue infiltration. The integrin-associated protein (CD47), expressed on neutrophils and endothelial and epithelial cells, is also involved in the transendothelial and transepithelial migration of neutrophils.4 Metalloproteinases may be involved in cleavage of L-selectin, allowing neutrophil migration through the basement membrane.

Absence of CD18 causes lack of CD18/CD11 heterodimers and is called leukocyte adhesion deficiency type 1 (LAD1). Neutrophils lacking CD18 roll normally along the endothelium but are unable to stick to the vessel wall or exit the circulation after chemotactic stimulation. Absence of LFA-1 (CD11a/CD18) makes neutrophils unable to bind tightly to and traverse activated endothelium to infected areas. Therefore, LAD1 patients have chronic neutrophil leukocytosis, partly from inability of neutrophils to bind tightly to endothelium and exit the circulation, thus leading to a reduction in the marginated pool and an increase in the circulating pool of neutrophils. Poor neutrophil penetration to sites of bacterial invasion leads to necrotic ulcers that lack neutrophils on biopsy. Absence of Mac-1 (CD18/CD11b or CR3) leads to inability to perform complement-mediated phagocytosis, although antibody-mediated phagocytosis remains intact.

LAD1 is diagnosed by fluorescent-activated cell sorting (FACS), which shows levels of CD18 and its coexpressed molecules CD11a, CD11b, and CD11c. Severe LAD1 (<0.5% of normal protein expression) is a disorder manifested by delayed umbilical stump separation, umbilical stump infection, persistent leukocytosis in the absence of active infection (>15,000/μL), and severe, destructive periodontitis with the loss of teeth and alveolar bone. Recurrent infections of the skin, upper and lower airway, bowel, perirectal area, and septicemia are common and usually due to Staphylococcus aureus or Gram-negative rods. Severe LAD1 patients should receive bone marrow transplantation in early childhood. Moderate LAD1 (2.5%–10.0% of normal protein expression) patients tend to be diagnosed later in life and have fewer life-threatening infections. Leukocytosis, delayed wound healing, and periodontal disease are still common.

Neutrophil “rolling” along the endothelium is mediated through selectins, surface glycoproteins on the endothelium, and sialyl-Lewis X (CD15s), a surface glycoprotein on neutrophils (see Fig. 30-1). Endothelial cells express e-selectin (CD62E) and P-selectin (CD62P), whereas leukocytes express L-selectin (CD62L). Although endothelial e-selectin and P-selectin bind to the sialyl-Lewis X (CD15s) antigen on neutrophils, the neutrophil molecule L-selectin probably binds to distinct antigens on endothelium, including CD34, and is highly sensitive to glycosylation. L-selectin is shed by neutrophils on activation, thereby allowing neutrophil migration into sites of inflammation. Cross-linking of L-selectin on neutrophils results in superoxide generation and may result in upregulation of TNF-α, IL-8, and tyrosine phosphorylation and activation of MAP kinase. LAD2 is the disease that occurs when CD15s is improperly fucosylated. These patients have neutrophilia, recurrentpulmonary, periodontal, and cutaneous infections, and abnormal chemotaxis. The defect is in the gene SLC35C1, a GDP-fucose transporter. Because the disease affects sugar transport it is now classified as a congenital disorder of glycosylation (CDG), and known as CDG IIc. Interestingly, although infections are common early in life, LAD2 (CDG IIc) patients appear to improve with age. Administration of oral fucose to some patients with LAD2 has been inconsistently therapeutic. Most patients also have mental retardation, short stature, distinctive facies, and the Bombay (hh) blood phenotype, indicating the multiple systems in which fucosylation is critical. LAD3 (previously known as LAD1 variant) is associated with a syndrome like Glanzmann's thrombasthenia, and is due to mutations in KINDLIN3 (FERMT3), a molecule responsible for β1, β2, and β3 integrin activation in leukocytes and platelets leading to recurrent infections and bleeding.

The CD18 integrin pathway is critical for inflammation in the skin, but not necessary for accumulation of neutrophils in the lung or peritoneum. Therefore, although CD18-dependent pathways are critical for cutaneous inflammation, CD18-independent pathways exist for pulmonary and peritoneal inflammation. Further, in the setting of congenital absence of CD18, compensatory pathways exist in the mouse to respond to peritoneal inflammation.

Phagocytosis

Phagocytosis is the culmination of object recognition, binding, signaling, adherence, cytoskeletal remodeling, engulfment, and membrane fusion. Two mechanisms are well characterized, one mediated by immunoglobulin and the other mediated by complement.

Receptors for the Fc portions of immunoglobulin G [IgG (FcγR)] are present on many components of the cellular immune response, including neutrophils, monocytes, macrophages, eosinophils, and basophils. FcγRI (CD64) is a receptor for IgG1 and IgG3 on monocytes, macrophages, and eosinophils and is upregulated on neutrophils after interferon-γ (IFN-γ) stimulation. FcγRII (CD32) binds IgG with rather low affinity, and prefers IgG1 and IgG3. FcγRIII (CD16) binds IgG1 and IgG3 with intermediate affinity. Neutrophil FcγRIIIB is bound to the membrane through a glycan phosphatidyl inositol linkage, which is largely cleavable by phosphatidyl inositol-specific phospholipase C. As FcγRIIIB has no cytoplasmic domain, its role in signal transduction may be through association with FcγRII. Cross-linking of these receptors by antibody leads to rapid engulfment of targets with the release of granule contents and oxygen metabolites into the phagolysosome. If the target is too large, degranulation occurs against the antibody-coated surface.

The complement receptors CR1 and CR3 are expressed on the surfaces of neutrophils, eosinophils, and basophils. Complement receptor 1 is designated CD35 and is found on many cell types. It binds C3b and enhances its degradation to C3dg by factor I, thereby removing that molecule from further activation of the alternative pathway. CR3, CD11b/CD18, binds iC3b and fibrinogen as well as certain bacteria, parasites, and fungi. CR1 and CR3 are not per se able to stimulate phagocytosis, but in the presence of a second signal, such as one given through FcγR or by cytokines, phagocytosis proceeds.

Mechanisms of Killing

Neutrophil granules contain enzymes and proteins for killing ingested bacteria and fungi. Some of these bactericidal mechanisms are dependent on the generation of oxygen metabolites for microbicidal activity, but others are not. In addition to mobilizing their own resources, neutrophils produce a multitude of cytokines that stimulate and attract other phagocytes as well as lymphocytes.

Oxygen-Independent Pathways

Bactericidal/permeability-increasing protein (BPI) is a highly potent antibacterial granule protein synthesized and stored in the primary granules. It is a highly basic (isoelectric point >9.6) protein of 452 amino acids and approximately 58 kDa. Sequence homology to lipopolysaccharide (LPS)-binding protein, a critical endotoxin binding acute-phase reactant, suggests that it acts by directly binding to LPS. BPI is cytotoxic to Gram-negative bacteria at concentrations as low as 10−9 M, but much less effective against Gram-positive organisms. Binding to LPS leads to insertion of BPI into the outer membrane of the organism and eventual insertion into the inner membrane. Arrest of bacterial growth is solely dependent on the N-terminal half of the molecule. The C-terminal fragment serves as an anchor to the membrane. BPI appears to act inside the phagolysosome. Not all Gram-negative rods are sensitive to BPI, especially Burkholderia (Pseudomonas) cepacia and Serratia marcescens, pathogens in patients who lack oxidative killing.

Defensins are small (<4 kDa) cationic proteins in the primary granules of neutrophils involved in killing ingested Gram-positive and Gram-negative bacteria. Defensins are synthesized during the promyelocyte/myelocyte stage as prepropeptides, stored predominantly in a dense subset of the primary granules as propeptides, and released during neutrophil degranulation. Defensins prefer an actively metabolizing target and can kill transformed mammalian cells as well as prokaryotes and yeasts. Defensins and BPI act synergistically against Gram-negative bacteria. Defensins are reduced in CHS but absent in specific granule deficiency.

Proteinase 3 is the antigen against which the antineutrophil cytoplasmic antibody is directed in Wegener granulomatosis (see Chapter 164). It is found predominantly in primary granules but is also in secondary and secretory granules. Synthesis and surface display are upregulated by cytokines such as TNF-α. Elastase is a myeloid serine protease in primary granules that has remarkable roles in host defense. Mice with elastase defects have decreased resistance to Gram-negative bacteria. However, human neutrophil elastase deficiency causes the syndromes of severe congenital neutropenia and cyclic neutropenia. Phospholipase A2 is also found in neutrophil granules, where its various isoforms act both directly and synergistically with BPI to kill intracellular bacteria.

Lactoferrin is an iron-binding protein present in specific granules and in mucosal secretions. When released into the phagolysosome, lactoferrin binds iron and inhibits the growth of phagocytosed bacteria and some fungi. The mechanisms through which lactoferrin exerts its antimicrobial action probably include the depletion of iron from an organism's environment, but iron binding-independent antimicrobial activities and direct immunomodulatory effects are also reported. Lactoferrin is readily released into the circulation after burns and experimental endotoxemia, presumably from neutrophil degranulation.

Oxygen-Dependent Pathways

Neutrophils can produce toxic oxygen metabolites through the reduced nicotinamide dehydrogenase phosphate (NADPH) oxidase complex, which typically is assembled in the wall of the phagolysosome. The NADPH oxidase catalyses the addition of an electron to molecular oxygen, leading to the formation of superoxide anion. Superoxide in turn is converted to hydrogen peroxide by superoxide dismutase. Inside the phagolysosome, the primary granule component myeloperoxidase converts hydrogen peroxide to hypohalous acid by the addition of a halogen (chloride in neutrophils forms bleach, bromide in eosinophils). The NADPH oxidase is a multiprotein complex, which is maintained in separate membrane-bound and cytosolic compartments. On cell activation, the cytosolic components translocate to the phagolysosome membrane, resulting in an active NADPH oxidase, which can produce the respiratory burst.

The components of the NADPH oxidase are named phox (phagocyte oxidase) proteins. p22phox and gp91phox are the α and β chains, respectively, of the cytochrome b558 complex, which resides in the wall of the secondary granule. The cytosolic compartment contains three factors, p47phox, p67phox, p40phox and the small guanine nucleotide (guanine triphosphate) binding protein rac. Assembly of the NADPH complex is caused by diverse stimuli. p47phox and p67phox contain src-homology type 3 domains (SH3 boxes), which bind to proline-rich targets in themselves and in other members of the NADPH complex. Stimulation leads to structural changes in p47phox that promote its interaction with p22phox through p47phox SH3 domains and p22phox C-terminal proline-rich sequences.

Pathologic mutations in the five required components impair the generation of phagocyte superoxide and cause CGD, a disease characterized by recurrent life-threatening infections with bacteria and fungi and exuberant granuloma formation (see Chapter 143).5

Pharmacologic Manipulation

When neutrophils are overrecruited into neutrophilic dermatoses, they can produce reactive oxygen intermediates, activate proteinases, and release chemotactic cytokines, which can contribute to tissue injury and inflammation. Therefore, it makes sense to aim treatment to suppress the generation of reactive oxygen intermediates, to inhibit neutrophil adhesion and chemotaxis, and to suppress the release of lysosomal enzymes and chemotactic factors.

Dapsone is often used for chronic neutrophilic dermatoses such as dermatitis herpetiformis, subcorneal pustular dermatosis, or erythema elevatum diutinum (see Chapter 225). Dapsone suppresses neutrophil adherence and subsequent migration, is a successful scavenger of reactive oxygen intermediates, and interferes with the myeloperoxidase-halide system and leukocyte-mediated cytotoxicity. Thalidomide reduces inflammation through inhibition of TNF-α and subsequent neutrophil–endothelial adhesion and reactive oxidant generation (see Chapter 235). Dapsone and thalidomide both inhibit the proinflammatory cytokine TNF-α, which activates neutrophils through upregulation of complement receptors (CR3, CR4) and endothelial adhesion molecules, ICAM-1 and e-selectin. Colchicine inhibits neutrophil chemotaxis, release of lysosomal enzymes, and production of reactive oxidants and is therefore considered for Sweet syndrome and neutrophilic bullous dermatoses. Antibiotics such as tetracyclines, macrolides, and metronidazole also have antioxidant properties and interfere with neutrophil chemotaxis. In addition to its antioxidant activity, sulfasalazine induces neutrophil apoptosis and enhances adenosine release at sites of inflammation, making it especially useful in pyoderma gangrenosum.

Ontogeny and Development

Eosinophils develop in the bone marrow from multipotential, stem cell-derived CD34+ myeloid progenitor cells in response to eosinophilopoietic cytokines and growth factors (see Fig. 31-1). They are released into the circulation as mature cells.1–3 Important stimulatory cytokines and growth factors for eosinophils include interleukin (IL)-3, granulocyte macrophage colony stimulating factor (GM-CSF), and IL-5. Activated T cells likely are the principal sources of IL-3, GM-CSF, and IL-5 that induce eosinophil differentiation in bone marrow. However, depending on pathogenic stimuli, eosinophilopoietic cytokines may be released by other cell types, including mast cells, macrophages, natural killer cells, endothelial cells, epithelial cells, fibroblasts, and even eosinophils, themselves.4 IL-3 and GM-CSF are pluripotent cytokines that have effects on other hematopoietic lineages. IL-5 is the most selective eosinophil-active cytokine, but it is relatively late acting. Although it is both necessary and sufficient for eosinophil differentiation, IL-5 demonstrates maximum activity on the IL-5 receptor (IL-5R)-positive eosinophil progenitor pool that first is expanded by earlier acting pluripotent cytokines such as IL-3 and GM-CSF4; expression of the high affinity IL-5R is a prerequisite for eosinophil development. Exodus from the bone marrow also is regulated by IL-5. IL-3, GM-CSF, along with IL-5, promote survival, activation, and chemotaxis of eosinophils through binding to receptors that have a common β chain (CD131) with IL-5R, and unique α chains. See ^ch31etb0.1 for designations of many factors involved in eosinophil biology.

Eosinophil/Basophil Colony Forming Unit

Corroborating Mouse Studies on Cytokines

Interactions of Eosinophilic Factors and Cytokines and Intracellular Signaling

The interactions of eosinophilopoietic factors with their receptors stimulate a cascade of complex biochemical events through signal transduction. Signaling events progress in four steps: (1) juxtamembranous signaling in which membrane-anchored tyrosine kinases and lipid kinases are activated; (2) signal interfacing which serves to transduce juxtamembranous signals to cytosolic signals; (3) mobile signaling in which cytosolic signaling molecules translocate from the receptor site to other cellular compartments including the nucleus, mitochondria, and cytoskeleton; and (4) transcription activation resulting from nuclear translocation and initiation of gene transcription. Studies have shown the pivotal role of IL-5 in immune responses involving eosinophils through receptor-driven signaling.20 IL-5 binds to the α chain of the IL-5R and induces recruitment of the common β (βc) chain to IL-5R. Janus kinase (JAK)2 tyrosine kinase is constitutively associated with IL-5Rα, and JAK1 tyrosine kinase with IL-5Rβc; both are activated with IL-5 binding as part of the juxtamembranous step. Lyn and Fes are other tyrosine kinases involved in the first step; these tyrosine kinases also are activated by IL-3 and GM-CSF. Adaptor proteins, src homologues and collagen (Shc), SH2-containing phosphatase-2 (SHP-2), growth factor receptor-bound protein 2 (Grb2), Vav, and lipid kinases, phosphatidylinositol 3-kinase (PI-3K), function in the interfacing step. The activation of JAK2 and signal transducer and activator of transcription (STAT) 5 is essential for IL-5 dependent signal transduction. The Ras GTPase-extracellular signal-regulated kinase (Ras-ERK) and also known as Ras-mitogen-activated protein kinase (Ras-MAPK) pathway, in addition to the JAK2-STAT5 pathway, is important in IL-5 signaling in the mobile step. The JAK-STAT and Ras-MAPK pathways converge at various levels in IL-5 signaling of eosinophils. IL-5 induces the expression of cytokine-inducible SH2 protein (CIS) and JAK-binding protein (JAB), which is one of the negative feedback loops in the regulation of IL-5 signaling. Multiple other interactive signal transduction pathways induce and regulate gene expression for eosinophil growth, development, activation, and survival.21,22 Much of the discovery in these pathways has been in murine systems with presumed general applicability to humans. However, at least part of the immune response to IL-5 in mice is NOT part of the biological effect in humans, i.e., in mice, IL-5 enhances several functions of B cells, including immunoglobulin production, growth, and differentiation, whereas human B-cells are influenced by IL-5 only in the presence of specific cytokines and under certain conditions.20 However, human IL-5 does act on T cells by increasing the expression of IL2Rα and augmenting cytotoxic T cell generation.23

Role of Transcription Factors

Eosinophil Ultrastructure and Granule Content

(See Fig. 31-2)

Mature eosinophils are 12–17 μm in diameter and, therefore, slightly larger than neutrophils. They typically have a bilobed nucleus with highly condensed peripheral chromatin. Eosinophils have distinctive cytoplasmic granules, demonstrated by their staining properties with acidic dyes such as eosin and by their unique electron microscopic appearance. These specific or secondary granules are composed of an electron-dense core and a less electron dense matrix, the core being a crystalline lattice by electron microscopy. In cross section, the eosinophil contains approximately 30 of these membrane-bound, core-containing, secondary granules.1 Five highly basic proteins are found within the granules: (1) major basic protein (MBP)-1, (2) MBP-2, (3) eosinophil-derived neurotoxin (EDN) also known as ribonuclease (RNase)2, (4) eosinophil cationic protein (ECP) also known as RNase3, and (5) eosinophil peroxidase (EPO). Several other types of proteins are found in secondary granules and include enzymes, cytokines, growth factors, and chemokines. Eosinophils contain three other types of cytoplasmic granules, referred to as (1) primary granules, (2) small granules, and (3) secretory vesicles. Primary granules are of variable size, round, uniformly dense, present in 1–3 per electron microscopic cross section, and more common in immature eosinophilic promyelocytes. These granules may contain Charcot–Leyden crystal protein (also known as galectin-10), which can be found in neutrophils as well31; Charcot–Leyden crystals (CLCs) are characteristically found in asthmatic sputum and in feces from patients with helminth infections or eosinophilic gastroenteritis. Small granules contain acid phosphatase and arylsulfatase and are present at 2–8 per electron microscopic cross section. Secretory vesicles, also referred to as tubulovesicular structures or microgranules, are characterized by their small, dumbbell-shaped appearance and their albumin content. They are the most abundant granules in number, with approximately 160 per electron microscopic cross section. Normal eosinophils contain varying numbers of nonmembrane-bound lipid bodies, which are the principal stores of arachidonic acid. Lipid bodies also contain the enzymes, cyclooxygenase, 5- and 15-lipoxygenase, which are required to synthesize prostaglandins, leukotrienes (LTs), and eoxins (vide infra), and are increased in activated eosinophils.1

Biologic Functions

Non-Mammalian Studies

Mammalian Studies

In mammals, such as the mouse and humans, the eosinophil is released as a mature cell into the circulation from the bone marrow, but is present in the blood only transiently, ranging from 8–18 hours. Eosinophils comprise a small portion, normally 6% or less, of circulating leukocytes. They are primarily tissue dwelling cells, but only in certain tissues in humans, with an average tissue life span of 2–5 days. This may be prolonged by cytokines that increase eosinophil survival for up to 14 days. Under normal circumstances, a balance exists between bone marrow production and release of eosinophils, their time in circulation, and their entrance into tissues. Changes in any one of the compartments causes an increase or decrease in circulating and tissue eosinophils. Eosinophilia in blood or tissue or both is associated with helminthiasis, allergic hypersensitivity, and other pathological conditions. In humans, bone marrow, spleen, lymph node, thymus, and gastrointestinal tract from the stomach through the colon, sparing the esophagus, are the only tissues in which eosinophils normally reside.38 Furthermore, the gastrointestinal tract is the only organ other than bone marrow in which extracellular eosinophil granule protein deposition is observed even under homeostatic conditions. Eosinophils and their granule proteins are found in the lamina propria in normal gastrointestinal tract and are not found in Peyer's patches or epithelium. Eosinophils may be important for thymocyte deletion based on the localization of eosinophils within the thymus and the timing of their migration during the neonatal period.39 Murine observations indicate that eosinophils are also important for postnatal mammary gland and uterine development, and their recruitment into the uterus heralds estrus. Although eosinophils, themselves, are not known to participate in human reproduction, eosinophil proMBP-1 is expressed in the uterus by placental X and giant cells during pregnancy, and its production peaks 2–3 weeks before parturition.40,41 The recruitment of eosinophils to the gastrointestinal, thymic, uterine, and mammary tissues is under the control of the CC chemokine, CCL11.42,43

Once eosinophils enter tissues, most do not recirculate. Several possible mechanisms exist for removal of tissue eosinophils. These include shedding of the cells across mucosal surfaces into the lumen of the intestinal or respiratory tract, engulfment of apoptotic eosinophils by macrophages, and lysis or degranulation with cellular degeneration. In various inflammatory conditions, including those affecting the skin (Chapter 36), striking numbers of free granules and/or eosinophil granule protein deposition are present in the absence of intact eosinophils.1 Studies recently have revealed that isolated eosinophil granules express extracellular domains for interferon (IFN)-γ receptor and CCR3 and, upon stimulation, respond independently as organelles by releasing ECP.44

Role of Eosinophils in Immune Function

Shortly after their discovery by Paul Ehrlich in 1879, eosinophils were observed in association with helminth infections. Theories have been promulgated that eosinophils are important for host defense against parasites spawning numerous studies.45 For example, in vitro studies demonstrated that eosinophils are cytotoxic to large nonphagocytosable organisms, such as multicellular helminthic parasites. Eosinophils bind to host-derived immunoglobulins and complement components on the surface of their targets (so-called antibody- or complement-) dependent cytotoxicity. They also bind to carbohydrate ligands expressed on parasites, such as the Lewisx-related molecules, and cell adhesion molecules similar to selectins. Eosinophils are activated to release their granule products with deposition of these biologically active proteins in and around the parasites causing disruption of the parasite's integument and, ultimately, death of the organism. The granule proteins have different effects. ECP produces fragmentation and disruption whereas MBP-1 produces a distinctive ballooning detachment of the tegumental membrane, and EDN is active only at high concentrations causing crinkling of the tegumental membrane.46 However, in murine models in which blood, marrow, and tissue eosinophilia is largely abolished by neutralizing IL-5 activity, the intensities of primary or secondary parasitic infection are unchanged indicating that eosinophils have little or no role in parasitic host defense in these models.1 The results must be interpreted cautiously because mouse and human eosinophils have functional differences, and mice are not natural hosts of many of the parasites tested experimentally.

Eosinophils also release cytotoxic granule proteins onto the surface of fungal organisms and into the extracellular milieu in fungal infections. Eosinophils kill fungi in a contact-dependent manner. Eosinophils adhere to the fungal cell wall component, β-glucan, via a β2-integrin surface molecule, CD11b.47 Eosinophils do not express other common fungal receptors, such as dectin-1 and lactosylceramide, and, specifically, do not react with chitin. However, chitin, which is a polymer that confers structural rigidity to fungi, helminths, crustaceans, and insects, induces accumulation of eosinophils in tissues through production of LT B4 in mice.48 Eosinophils also are activated by fungal organisms that release proteases, such as Alternaria, through protease-activated receptors (PARs). For example, fungal aspartate protease activates eosinophils through PAR-2 and, thereby, mediates eosinophils’ innate responses to certain fungi.49

As a granulocyte, the eosinophil is capable of phagocytosing and killing bacteria and other small microbes in vitro, but eosinophils cannot effectively defend against bacterial infections when neutrophil function is deficient. Nevertheless, recent investigations reveal that eosinophils may have a role in innate immunity against bacteria using a unique mechanism, DNA trap.50 Eosinophils rapidly release mitochondrial DNA when exposed to bacteria, a complement component, C5a, or CCR3 ligands. The traps contain eosinophil granule proteins, ECP and MBP, and have antimicrobial effects. In the extracellular space, the granule proteins and mitochondrial DNA form structures that bind and kill bacteria both in vitro and in vitro. Eosinophils, unlike neutrophils, do not undergo cell death as part of this process. This may be an important innate immune response, particularly in mucosal epithelium.50

Another protective function that eosinophils may have is in viral infections. Eosinophils and their granule proteins are increased in the respiratory tracts of patients with respiratory syncytial virus, an RNA-viral infection. EDN (RNase2) and ECP (RNase3), eosinophil granule matrix proteins, are ribonucleases (vide infra). Purified eosinophils, as well as EDN and ECP individually, reduced viral titers when added to respiratory syncytial viral suspensions. In mice, at least 11 eosinophil-associated ribonucleases degrade single stranded RNA containing viruses.51 Despite divergence of the coding regions, conserved eosinophil ribonuclease activity across species suggests a strong evolutionary pressure to preserve this critical enzymatic activity.51 In other studies, pretreatment of parainfluenza-infected guinea pigs with anti-IL-5, to reduce numbers of eosinophils, strikingly increased viral load in the airways. Viruses, including parainfluenza virus, respiratory syncytial virus, and rhinovirus, induce the release of another eosinophil granule protein, EPO, when coincubated with antigen-presenting cells and T cells.52 Paradoxically, eosinophils may be a potential reservoir for the human immunodeficiency virus (HIV)-1.53

Eosinophils may have other roles in immune responses as well. Through MHC class II expression and IL-1α production, they can function as antigen-presenting cells for a variety of viral, parasitic, and microbial antigens, including staphylococcal superantigens, and allergens.54,55 Eosinophils are recruited to secondary lymphoid structures to promote the proliferation of effector T cells although they are unable to affect naïve T cells.56 Eosinophils, as sources of cytokines, influence T cell-dependent responses.1 In keeping with the prominence of eosinophils in allergic disorders, eosinophils are involved in T cell polarization favoring Th2 by promoting Th1 apoptosis in addition to their influence via cytokine expression.54,57–60

Role of Eosinophils in Disease

The activities of eosinophil-derived products include direct cytotoxic effects on structural cells and microbes, increased vascular permeability, procoagulant effects, innate immune responses to some parasites, viruses, fungi and tumor cells, enhancement of leukocyte migration, amplification of effector T-cell responses and, possibly even mammary gland development. Collectively, these varied biologic actions provide the pathophysiological basis for the signs and symptoms observed in eosinophil-associated diseases.

Eosinophils in lymph nodes and spleen are especially increased after allergen exposures or microbial insults.61,62 Eosinophils have been found in several cancers, particularly in lymphomas, leukemias, and colon cancer. Clinical studies indicate that certain tumors associated with tissue and/or peripheral eosinophilia have a more favorable prognosis,63 whereas in other tumors, they are thought to confer a poor prognosis, such as nodular sclerosing Hodgkin disease, Sézary syndrome, and gastric carcinomas. In Sézary syndrome (see Chapter 145), the tumor cells produce IL-5 and, therefore, are responsible for the eosinophilia, which is a reflection of tumor burden.64 Where eosinophilia is a good prognostic factor, eosinophils are considered to be part of an effective host response to the tumor.65,66

Antitumor Activity, Gastrointestinal Diseases and Asthma

Eosinophil Constituents and Their Activities

The eosinophil contains and produces a myriad of factors that implicate its role in inflammation and tissue destruction and remodeling (see Fig. 31-2).70 Products released by eosinophils include chemoattractants, colony-stimulating factors, and endothelial-activating cytokines. In addition to toxic cationic proteins from specific granules and oxidative products released into tissues following activation, these factors include arachidonic acid-derived lipids, hydrolytic enzymes, neuropeptides, colony-stimulating factors, and cytokines/chemokines that facilitate further leukocyte recruitment to sites of inflammation (see Fig. 31-2). Surface molecule expression is important in all aspects of eosinophil biology from promoting growth and differentiation to eosinophil trafficking into tissue to activation and/or priming of the cells to senescence. Numerous membrane factors are expressed on eosinophils that further direct eosinophil biological effects. (See ^ch31etb0.1 for designations of various eosinophil factors.)

Eosinophil Granule Proteins

Among the products of eosinophils that are most damaging to the host are the specific granule's cationic proteins. Furthermore, the granule proteins are markers of eosinophil activity because the eosinophil often loses its characteristic morphology through cytolysis in tissues.71 Knowledge of their biological actions provides insight into their functions in human disease. Once deposited, the granule proteins persist in tissues for extended times—EPO for 1 week, ECP for 2 weeks, EDN for 2.5 weeks, and MBP-1 for 6 weeks.72 Each of these proteins have been shown to induce direct tissue damage to both host cells, including myocytes, endothelium, neurons, epithelium, and smooth muscle, and to microbes (vide supra). For example, MBP-1, ECP, or EPO application to airway epithelium in primates produces ciliostasis, desquamation, and hyperreactivity of respiratory smooth muscle mimicking the pathology of asthma.73,74 Damage to endothelium in eosinophilic endomyocardial disease is thought to be the initiating factor in the cardiomyopathy observed in the hypereosinophilic syndromes (see Chapter 36).75 All four of the cationic granule proteins [(1) EPO, (2) ECP, (3) EDN, and (4) MBP-1] likely contribute to the edema observed in skin diseases due to their vasodilatory effects with contribution from mast cells and basophil histamine release by MBP-1.76 Eosinophil granule proteins stimulate various cells in addition to mast cells and basophils, including neutrophils and platelets. Nodules, eosinophilia, rheumatism, dermatitis and swelling (NERDS), episodic angioedema with eosinophilia (Gleich syndrome), urticaria, eosinophilic cellulitis (Wells syndrome), and insect bite reactions demonstrate variable degrees of edema that are probably explained, at least in part, by this mechanism (see Chapter 36). Eosinophil granule proteins injected into skin produce lesions including dose-dependent wheal-and-flare reactions by MBP and ulcerations by ECP and EDN.77,78 Wound healing is delayed in the presence of eosinophils and eosinophil granule proteins.79,80 Eosinophils, through activities of granule proteins, have procoagulant activity. Thromboses have developed in the hypereosinophilic syndromes, including several case reports of hepatic vein obstruction (Budd–Chiari syndrome). The thromboses could be the result of direct endothelial damage or due to the ability of MBP and ECP to neutralize heparin. In addition, MBP is a strong platelet agonist, and platelet-activating factor (PAF), which is released by eosinophils, causes platelet aggregation.1

Major Basic Protein

General Characteristics

Major basic protein (MBP) comprises the crystalloid core of the specific eosinophil granule. It was so named because it accounts for a major portion, about 55% (in guinea pig), of the eosinophil granule protein, and has a high isoelectric point, calculated at greater than pH 11, so strongly basic that it cannot be measured accurately. It is now known that MBP is expressed as two homologs, MBP-1 and MBP-2, coded by different genes on chromosome 11.

Detailed Major Basic Protein's Chemistry

Tissues Effects of Basic Proteins

MBP-1 directly damages helminths and also lethally damages mammalian cells and tissues, examples of which are its ability to cause exfoliation of bronchial epithelial cells and to kill tumor cells. MBP-1 exerts its effects by increasing cell membrane permeability through surface charge interactions leading to disruption of the cell surface lipid bilayer.81 MBP-1 and MBP-2, but none of the other eosinophil granule proteins, stimulate histamine and LTC4 release from human basophils. Further, MBP-1 and MBP-2 stimulate neutrophils, inducing release of superoxide, lysozyme, and IL-8. MBP-1 and EPO are potent platelet agonists causing release of 5-hydroxytryptamine and promoting clotting.

Eosinophil Peroxidase

Eosinophil peroxidase (EPO) is highly basic, pI 10.8, localized in the matrix of the specific eosinophil granule and is a key participant in generating reactive oxidants and free radical species in activated eosinophils. EPO consists of a heavy chain and a light chain encoded with a prosequence. The EPO gene is on chromosome 17 and maps closely to myeloperoxidase and lactoperoxidase genes, two other members of the mammalian peroxidase family found in neutrophils and mucosal glands, respectively. Although MBP is present in the highest molar concentration in eosinophil granules, EPO, by weight, is the most abundant protein constituting approximately 25% of the specific eosinophil granule's total protein mass. EPO catalyzes the oxidation of halides, pseudohalides, and nitric oxide to form highly reactive oxygen species (hypohalous acids), reactive nitrogen metabolites (nitric dioxide), and other oxidants that then oxidize targets on proteins with oxidative stress and subsequent cell death by apoptosis and necrosis. EPO kills numerous microorganisms in the presence of hydrogen peroxide, generated by eosinophils and other phagocytes, and halide. This combination of products also initiates mast cell secretion. As noted, both EPO and MBP induce noncytolytic, dose-dependent 5-hydroxytryptamine release from platelets.82 Binding of EPO to neutrophils reversibly inhibits EPO peroxidase activity but increases neutrophil aggregation and adhesion to endothelial cells.83 EPO binding to microbes, including Staphylococcus aureus, greatly potentiates their killing by phagocytes. EPO-coated tumor cells are spontaneously lysed by activated macrophages.

Eosinophil Cationic Protein and Eosinophil-Derived Neurotoxin

General Characteristics

Eosinophil cationic protein (ECP or RNase3) and eosinophil-derived neurotoxin (EDN or RNase2) are homologous proteins with sequence identity in 37 of 55 amino acid residues and are encoded on chromosome 14. ECP also has neurotoxic activity. ECP and EDN play a role in viral host defense to RNA viruses.51,88,89 New roles for these proteins continue to be identified.90 EDN induces the migration and maturation of dendritic cells.91 It also is an endogenous ligand of Toll-like receptor 2 (TLR2) and can activate myeloid dendritic cells by triggering the TLR2-myeloid differentiation factor 88 (Myd88) signaling pathway.60 Based on its ability to serve as a chemoattractant and activator of dendritic cells along with enhancing antigen-specific Th2-biased immune responses, EDN functions as an alarmin alerting the adaptive immune system to preferentially enhance antigen-specific Th2 responses.60

Specific Chemicals Characteristics of ECP and ECN

They are found in the matrices of specific granules in mature eosinophils. Both proteins also are present in mature neutrophils.84 They are highly basic proteins, the pI of ECP is 10.8 and the pI of EDN is 8.9. It is likely that these two proteins arose as a consequence of gene duplication 25–40 million years ago.85 EDN, in particular, is a rapidly evolving protein, accumulating nonsilent mutations at a rate exceeding those of most other functional coding sequences studied in primates.86 Although the homology in EDN and ECP sequences accounts for certain similarities, their differences are considerable. ECP and EDN are members of the human RNase A family, ECP is a relatively weak ribonuclease, and EDN is a potent ribonuclease. ECP is a potent toxin for parasites through a different mechanism than MBP-1 and is more effective at killing certain helminths than MBP-1. EDN, as its name implies, was discovered as the neurotoxin that accounted for the Gordon phenomenon, described in 1933, in which splenic extracts from patients with the Hodgkin's disease caused severe damage to myelinated neurons in experimental animals.87

Charcot–Leyden Crystal Protein

General

Distinctive hexagonal, bipyramidal crystals were initially described in 1853 in a patient with leukemia and later, in 1872, in the sputa of asthmatic patients. Since then, CLCs have been regarded as a hallmark of eosinophilia. CLC protein is an abundant, characteristic, although not unique, protein of eosinophils. It is also found in lesser amounts in basophils. CLC mRNA and EDN mRNA are among the most highly expressed mRNAs in mature peripheral blood eosinophils suggesting de novo synthesis.1

Charcot–Leyden Crystals

Specific Biochemistry

Other Eosinophil Enzymes

Reactive Oxygen Metabolites

Reactive oxygen species are important mediators of the tissue injury caused by eosinophils. The reactive oxygen intermediates produced by NADPH oxidase during the respiratory burst initiated on eosinophil activation include superoxide, hydroxyl radicals, and hydrogen peroxide. Hydrogen peroxide can generate hypohalous acids through the action of EPO, which are cytotoxic as well. Reactive oxygen species also can augment the inflammatory response by inducing gene expression and T-cell proliferation.98

Lipid Mediators

General

Lipid bodies in eosinophils are storage sites for arachidonic acids. Eosinophils produce several arachidonic acid metabolites, including cysteinyl LTs from the 5-lipoxygenase pathway (LTC4, LTD4, and LTE4) and thromboxanes and prostaglandins from the cyclooxygenase pathway [thromboxane B2, prostaglandin (PG) E2, and PGF1α].99,100

Detailed Lipid Mediator Biochemistry

Cytokines

Eosinophils are a considerable source of growth factors and regulatory as well as proinflammatory cytokines and chemokines.1,103 The various growth factors produced by eosinophils include transforming growth factor (TGF)-α, TGF-β, fibroblast growth factor-2, vascular endothelial growth factor, nerve growth factor, and platelet-derived growth factor-β. There is evidence that these growth factors induce stromal fibrosis and basement membrane thickening at sites of chronic eosinophilic inflammation including nasal polyps, asthmatic airways and, likely, in certain skin disorders such as atopic dermatitis.1 The production and release of nerve growth factor by eosinophils promotes neurites in nerve cells.104 Another group of cytokines produced by eosinophils modulates other immune cells and includes tumor necrosis factor (TNF)-α, macrophage inflammatory protein (MIP)-1α (CCL3), IL-1α, IL-2, IL-3, IL-4, IL-5, IL-6, CXCL8 (IL-8), IL-10, IL-12, IL-13, IL-16, GM-CSF, and IFN-γ.103 Additional chemokines produced by eosinophils are CXCL13 (B lymphocyte chemoattractant factor), CCL5 [regulated on activation, normal T cells expressed and secreted (RANTES)], and CCL11, in addition to CCL3 and CXCL8. CCL5 apparently is stored in at least two intracellular compartments, namely, (1) the matrix of specific eosinophil granules and (2) small secretory vesicles. All of these cytokines are constitutively produced in low levels in resting eosinophils and induced in inflammatory conditions with activation of eosinophils by engagement of receptors (vide infra) with immunoglobulins, complement and cytokines, including those produced by eosinophils, themselves, in an autocrine manner. Notably, eosinophils produce the three principal cytokines involved in their own growth and differentiation—(1) IL-3, (2) IL-5, and (3) GM-CSF, as well as chemokines important in their own chemotaxis, CCL5 and CCL11. It is known that eosinophils synthesize and secrete GM-CSF by a peptidyl-prolyl isomerase (PIN1)-dependent mechanism.105 However, although eosinophils produce quantities of certain cytokines comparable to T cell production,106 the relative contributions of eosinophil-derived cytokines to inflammation remain to be determined. In summary, the eosinophil-derived cytokines may function in both an autocrine and paracrine fashion and likely have pathophysiological relevance.

Surface Expression

Eosinophils express numerous receptors and other factors on their surface membranes through which they communicate with the extracellular environment, but no single surface protein is uniquely expressed on eosinophils. These receptors have been identified either by flow cytometry or by functional assays, and can be grouped as follows: chemotactic factor and complement receptors including chemokine, LT and PAF; immunoglobulin supergene family member receptors including immunoglobulins; cytokine receptors including those discussed above; adhesion molecule receptors; receptors involved in apoptosis; and miscellaneous receptors and surface factors. Eosinophil membrane proteins are promising targets for therapeutic modulation of eosinophil effects (see section “Pharmacological Manipulation”).

Chemotactic Factor and Complement Receptors

Chemotactic factors are important in orchestrating cellular trafficking to sites of inflammation as well as physiologic homing (e.g., eosinophils to gastrointestinal tract). The eosinophil has receptors for many chemotactic agents such as LTB4, PAF, bacterial products (N-formyl-methionyl-leucyl-phenylalanine), complement anaphylatoxins, C3a and C5a. Eosinophils express complement receptor (CR)1 (CD35), a receptor for C1q that also binds C4b, C3b, and iC3b, and CR3 (Mac-1, CD11b/CD18) in addition to receptors for C3a and C5a. These are important receptors in eosinophil effector functions. The binding of chemokines to their respective receptors mediates many biological effects, which, in addition to cell shape change and migration, includes cell activation, receptor internalization, induction of the respiratory burst with generation of toxic oxygen metabolites, and transient activation of integrin adhesiveness. PAF activity on eosinophils is mediated through PAF receptors that have been cloned. PAF is one of the most potent chemoattractants for eosinophils and selectively recruits eosinophils over neutrophils. PAF also induces release of granule proteins, reactive oxygen species, and LTC4 from eosinophils. LTB4 through its eosinophil receptor potently stimulates eosinophil chemotaxis and elicits arachidonic acid metabolism and a respiratory burst. It does not induce eosinophil degranulation. Eosinophils also may have LTD4 receptors.1 The chemotaxins listed above have potent effects on eosinophils but are nonselective in that they are active on other leukocytes. Because many eosinophil-associated diseases are characterized by tissue eosinophil infiltration with little or no neutrophil infiltration, the identification of the CCR3 receptor and its ligands was an important breakthrough in discovering eosinophil-selective chemotaxins.107 Specific members of the chemokine family are critical for the cellular trafficking of eosinophils. The major ligands of CCR3, CCL5, CCL11, CCL13 (monocyte chemotactic protein-4), CCL24 (eotaxin-2), CCL26 (eotaxin-3), play a critical role in both the homeostatic and inflammation-induced recruitment of eosinophils to tissue sites.13,108 Not surprisingly, the most highly expressed chemokine receptor on human eosinophils is CCR3 (40,000–400,000 receptors per cell). CCR3 is a seven-transmembrane spanning G-protein coupled receptor that can deliver both powerful positive and negative signals depending on the interacting ligand.109

Immunoglobulin Supergene Family Member Receptors

Many of the studies of eosinophil functions, including phagocytosis, antigen-dependent cytotoxicity, oxygen metabolism, LTC4 production, and eosinophil survival, have been performed using IgG-coated targets. Among eosinophil surface receptors for the immunoglobulin family members, the most highly expressed receptor is FcγRII (CD32), which binds aggregated IgG, particularly of the subclasses IgG1 and IgG3. The binding of IgG to this receptor may be important in eosinophil degranulation in parasitic and allergic diseases along with other eosinophil functions.1 Freshly isolated eosinophils express only FcγRII (CD32) of the IgG receptors, but eosinophils can be stimulated by cytokines, particularly IFN-γ, to express FcγRI (CD64) and FcγRIII (CD16) and augment FcγRII (CD32) expression.

FcαRI (CD89), the IgA receptor, is present on the surface of eosinophils. Eosinophils also possess a binding site for secretory component that may account for the finding that secretory IgA is the most potent immunoglobulin stimulant for eosinophil degranulation. The interaction of eosinophils with IgA is enhanced by Th2 cytokines, IL-4 and IL-5. These observations, coupled with the observation that many eosinophils are found in epithelial tissues such as gastrointestinal tract, suggest that eosinophils and secretory IgA play an important role in mucosal immunity.

Members of the immunoglobulin superfamily are type I transmembrane molecules that share common structural characteristics of the globular domains found in immunoglobulins. Intercellular adhesion molecule (ICAM)-1 (CD54) and ICAM-3 (CD50) are members of this superfamily expressed on eosinophils and are likely important in leukocyte–leukocyte and leukocyte–tissue cell adhesion through leukocyte function-associated antigen (LFA)-1 (αLβ2; CD11a/CD18) as its counterligand (see Fig. 31-1).

Cytokine Receptors

Cytokine receptors are present at low levels on the surfaces of eosinophils. Receptors for IL-3 (CD123), IL-5 (CD125), and GM-CSF (CD116) are readily detected and, as noted previously, all share a common β chain (CD132). Eosinophil activation has been observed by a variety of other cytokines through presumed and/or detected receptors. These include: stem cell factor (c-kit receptor; CD117), IFN-γ (CD119), TNF-α (CD120), IL-4 (CD124), IL-9 (CD129 and CD132), IL-13 (gp65), IL-2 (CD25), IL-31, and TGF-β receptors. Many of these receptors are for cytokines that eosinophils produce providing further evidence that they have autocrine functions.

Adhesion Molecule Receptors

Adhesion molecule receptors are expressed on the eosinophil cell surface to mediate trafficking to and within tissues, and for general cell–cell interactions.112 These receptors fall into three groups: (1) immunoglobulin superfamily (reviewed above), (2) selectins and their glycoprotein counterligands, and (3) integrins. l-selectin (CD62L) and p-selectin glycoprotein ligand-1 (PSGL-1, CD162) are expressed at high levels on eosinophils, whereas e-selectin ligands, as an example, sialyl-Lewis X (CD15s), are expressed at very low levels. p-selectin together with PSGL-1 is the most important selectin pair in eosinophil migration into tissues.

Eosinophils express a variety of integrins (β1, β2, and β7) on their surface, which facilitate their adhesion to extracellular matrix proteins, vascular cellular adhesion molecule (VCAM)-1 (CD106) on activated endothelium, or ICAM-1 present on resting or activated epithelium and activated endothelium. Integrins are composed of two subunits that exist as noncovalently associated heterodimers, with α and β subunits. The β1 integrins expressed on eosinophils include α4β1 (very late antigen-4 or VLA-4), which binds to VCAM-1 found on activated endothelium and the extracellular matrix protein, fibronectin. Eosinophil adhesion to fibronectin induces the autocrine production of eosinophil-activating survival cytokines, IL-3, IL-5, and GM-CSF. The only other β1 integrin expressed on eosinophils is α6β1, which mediates the binding to another matrix protein, laminin. Four β2 integrins are found on eosinophils: (1) αLβ2 (LFA-1), (2) αMβ2 (CD11b/CD18 or Mac-1), (3) αXβ2, and (4) αDβ2. These integrins bind to ICAM-1, ICAM-2, ICAM-3, VCAM-1, fibrinogen, and the complement fragment, C3bi. Lastly, eosinophils also express α4β7, which is the ligand for the gut mucosal addressin cell adhesion molecule-1 (MAdCAM-1), which is likely important in homing of eosinophils to gastrointestinal mucosa; α4β7 also mediates adhesion to fibronectin and VCAM-1.

Receptors Involved in Apoptosis

Eosinophils express several “death receptors,” that are involved in apoptotic pathways, such as Fas receptor (CD95), Siglec-8, CD30, CD45, Campath (CD52), and CD69, along with important intracellular regulators of eosinophil apoptosis such as the members of the B-cell leukemia/lymphoma (Bcl)-2 and inhibitor of apoptosis families.113 Diseases characterized by eosinophilia likely result, in part, from delayed or defective apoptotic pathways allowing accumulation and persistence of eosinophils in blood and/or tissues.

Miscellaneous Surface Receptors and Other Surface Factor Expression

Factors Working Together

The various products elaborated by eosinophils in response to receptor activation do not necessarily function independently but often act in concert to mediate their biological effects. For example, the release of TGF-α, TGF-β, fibroblast growth factor-2, vascular endothelial growth factor, MMP-9, and inhibitors of MMPs from activated eosinophils collectively induce fibroblast proliferation and extracellular matrix protein production. Eosinophils contribute factors of their own and influence factor production from other cells, for example, eosinophil mediators induce platelet release of TGF-β. After intradermal eosinophil infiltration, there is production of extracellular proteins, including tenascin and procollagen 1, as well as myofibroblast formation.120 Eosinophil-induced fibrosis is observed in the lungs and heart of patients with hypereosinophilic syndrome, in and around organs in other fibrosing/sclerosing disorders and in the skin of patients with eosinophilic fasciitis (Shulman syndrome), eosinophilia–myalgia syndrome, and toxic oil syndrome (see Chapter 36).121 Eosinophil granule proteins, MBP-1 and EDN, along with other neuroactive mediators produced by eosinophils such as nerve growth factor, vasoactive intestinal peptide, and substance P, likely affect nerve physiology. In fact, eosinophils and eosinophil granule proteins are often observed in close proximity to nerve endings.122,123 Eosinophil-induced nerve dysfunction is likely an important part of the gastric dysmotility observed in subjects with food allergies, the dysfunction of vagal muscarinic M2 receptors observed in asthmatics, and may also contribute to itch along with other physiological aberrations in atopic dermatitis and other cutaneous diseases.123,124 Collectively, the eosinophil's response to surface factors determines its role in health and disease.

Tissue Trafficking

The selective recruitment of eosinophils into sites of inflammation results from interactions among eosinophil-activating cytokines, chemokine-inducing cytokines, and endothelial-activating cytokines (see Fig. 31-1). Similar to other leukocytes, selectin, integrin, and immunoglobulin gene superfamily members contribute to the signaling involved in eosinophil trafficking. In particular, eosinophils constitutively express the integrin, VLA-4, which interacts with its ligand, VCAM-1, induced on endothelial cells by cytokines, especially Th2 cytokines (IL-4 and IL-13).125 After movement through vessels, eosinophils adhere to extracellular matrix proteins. Here, surface factors, b2 integrins such as CD11b/CD18 (Mac-1), bind to fibrous proteins such as fibronectin, laminin, and collagen, and, not only determine where eosinophils will reside, but likely prolong their survival.126 In this regard, CD11b/CD18 (Mac-1) integrin is also critical for eosinophil effector functions, including degranulation.127

Detailed Mechanisms of Tissue Recruitment

Eosinophil-Activating Cytokines

General

Eosinophil-activating cytokines can be produced by many cell types in addition to T cells and mast cells, including keratinocytes, endothelial cells, and monocytes, along with eosinophils, themselves. The eosinophil-activating cytokines, IL-3, IL-5, GM-CSF, and others, enhance chemotactic responses, in addition to multiple other effects on eosinophils such as promoting maturation, cell survival, and LT production.128

Specific Studies of Activating Cytokines

Endothelial-Activating Cytokines

(See Chapter 162)

During eosinophil migration, at least three types of endothelial activations occur. The first is the expression of p-selectin, which occurs when Weibel–Palade bodies in endothelial cells are transported to the cell surface rapidly after exposure to histamine, LTs, and a host of other inflammatory mediators. Expression of p-selectin on the endothelial-cell surface initiates leukocyte rolling, via CD162 (PSGL-1), which is the important initial step before firm adhesion and transendothelial migration. A second type of endothelial activation is that induced by nonspecific activators such as IL-1 and TNF-α. These cytokines stimulate endothelial expression of e-selectin, ICAM-1, and VCAM-1, to which eosinophils firmly adhere, or “tether.” They also induce production of chemokines by endothelial cells. The third type of endothelial activation is that induced by IL-4 and IL-13. These cytokines selectively induce VCAM-1, which is centrally involved in the recruitment of VLA-4-positive cells, including eosinophils, basophils, and lymphocytes into sites of allergic inflammation.

Chemokines

The transition from rolling to firm adherence is substantially increased by CCR3 ligands, the CC chemokines. Induction of the expression of chemokines by activated endothelial cells results in higher levels of chemokines on or near the endothelial surface, which transiently affect β1 and β2 integrin avidity, resulting in firm adhesion of the eosinophil to the endothelial cell. However, chemokines produced by structural cells such as fibroblasts, smooth muscle cells, and epithelium probably are more important in directing migration and activation of eosinophils within tissues than those expressed on endothelial cells.137

Chemokines as Eosinophil Chemoattractants

Tissue chemokine expression forms a gradient signal that guides eosinophils into tissue. CCL11 guides eosinophils into tissue locations in which eosinophils are normally present, thymus, uterus, mammary gland, and gastrointestinal tract.143 In Th2 disorders, Th2 cytokines induce chemokine expression. In lungs affected by asthma and eosinophilic pneumonia, CCL11, CCL24, and CCL26 are increased along with other chemokines,144 such as LTB4 and galectin-9.145,146 In the intestinal tract, CCL26 plays a key role in eosinophilic esophagitis, whereas CCL11 is involved in lower gastrointestinal eosinophil disorders. In skin, IL-4, IL-13, and TNF-α stimulate CCL11, CCL24, and CCL26 production from mast cell and lymphocyte sources as well as from fibroblasts (CCL11 and CCL26) and keratinocytes (CCL26).147 As in eosinophilic esophagitis, CCL26 may be important in atopic dermatitis in which serum CCL26 levels correlate with disease activity.148

Arachidonic acid metabolites, particularly, the cysteinyl LTs, LTC4, LTD4 and LTE4, and PGD2, are involved in eosinophil trafficking as evidenced by the observations that LT receptor antagonists reduce blood and lung eosinophilia and that mice, depleted of LTB4 receptors, have markedly reduced lung eosinophilia after allergen exposure. Eosinophil, basophil, and Th2 cell recruitment occurs, to some extent, through CRTH2 (CD294), the high affinity PGD2 type 2 receptor. Other factors that contribute to eosinophil trafficking are being identified. For example, eosinophils express high levels of histamine 4 receptors that mediate chemoattraction and activation.149 An extracellular matrix protein, periostin, encoded by an IL-13 inducible gene, likely facilitates eosinophil infiltration into tissues by directly affecting eosinophil adhesion; periostin is overexpressed in eosinophilic esophagitis and correlates with eosinophil numbers in biopsies.150

Activation of Eosinophils

As reviewed previously, various inflammatory mediators activate eosinophils. In addition to cytokines, TNF-α, GM-CSF, IL-3, and IL-5, these include complement components, C3a and C5a, lipid mediators, LTC4 and PAF, chemokines, as well as engagement of IgA and IgG Fc receptors. CD11b/CD18 (Mac-1)-dependent cellular adhesion is a critical component for degranulation and superoxide production induced by GM-CSF and PAF eosinophil activation and likely is an in vitro mechanism that results from eosinophil contact with stromal cells and/or proteins.151 Members of the CC chemokine subfamily, CCL5, CCL7 (MCP-3), CCL11, CCL13 (MCP-4), CCL24 that bind to the chemokine receptor, CCR3, also potently activate eosinophils. Activated eosinophils develop a number of phenotypic changes, including a reduction in granules, vacuolization, and an expansion of their cytoplasm, leading to a reduction in cell density and are referred to as hypodense. The number of hypodense cells predicts allergic disease severity. A cell surface marker that distinguishes hypodense from normodense eosinophils has not been identified, but there are several surface markers with enhanced expression on in vitro or in vitro activated or hypodense cells: αM integrin (CD11b), αX integrin (CD11c), FcγR111 (CD16), hyaluronic acid receptor (CD44), ICAM-1 (CD54), CD69, and HLA-DR.152

Upon recruitment and activation in tissues, eosinophils have various effects as detailed in previous sections. In tissues, eosinophils release granule contents into their extracellular space via three mechanisms: (1) piecemeal degranulation, (2) regulated secretion (also referred to as regulated exocytosis), and (3) cytolytic degranulation. First, in piecemeal degranulation, eosinophils selectively release specific granule components; as an example, IFN-γ activation promotes mobilization of granule-derived CCL5 to the eosinophil's surface without inducing cationic protein release.153–155 Second, in regulated secretion, a docking complex forms consisting of soluble N-ethylmaleimide-sensitive factor attachment protein (SNAP) receptors (SNAREs) located on the vesicle (vSNAREs) and the target membrane (tSNAREs). Two types of SNAREs have been described based on the presence of a conserved amino acid, arginine (R) or glutamine (Q). Human eosinophils express the R-SNARE, vesicle-associated membrane protein (VAMP)-2 on cytoplasmic secretory vesicles, and the Q-SNAREs, SNAP-23, and syntaxin-4, on the plasma membrane.156 VAMP-7 also plays a critical role in both eosinophil and neutrophil mediator release.157 The current understanding is that receptor-induced eosinophil activation leads to rapid mobilization of cytoplasmic vesicles to the plasma membrane, leading to the formation of a SNARE complex (VAMP-2, -7/SNAP-23/syntaxin-4) and mediator release. Third, cytolytic degranulation occurs in many inflammatory diseases including skin disease, such as atopic dermatitis, eosinophilic esophagitis, and lesions found in affected organs with the hypereosinophilic syndromes. It is characterized by organelle rupture, chromatolysis of nuclei with loss of morphologic integrity and identity of eosinophils, and extensive deposition of eosinophil granules and granule products in tissue.71 Therefore, it seems somewhat paradoxical that eosinophils from atopic dermatitis patients have prolonged eosinophil survival and yet exhibit such marked cytolytic degranulation in skin lesions. Clearly, there is much more to learn about eosinophil biology and its relevance to human diseases.

Pharmacological Manipulation

Eosinophil-associated disease is a term that, strictly speaking, refers to diseases in which eosinophil numbers or eosinophil granule protein levels (or other eosinophil products) are associated with disease activity. This term encompasses multiple heterogeneous disorders, including skin diseases (see Chapter 36), in which targeting eosinophils and/or their products is a therapeutic goal. Many available treatments reduce eosinophils numbers, thereby inhibiting eosinophilic inflammation, including glucocorticoids, calcineurin inhibitors, IFN-α, IFN-γ, LT antagonists, myelosuppressive/cytotoxic drugs, and, possibly, even antihistamines. However, none is specific for eosinophils. It is only in recent years that selective and direct reduction of eosinophils has been achieved, and these therapies have provided new insight into disease pathogenesis.9

Among the nonselective drugs for eosinophil reduction, glucocorticoids generally are very effective. The immediate (within 3 hours) reduction in circulating eosinophils observed after systemic administration of glucocorticoids likely occurs as a consequence of sequestration into extramedullary organs (liver, spleen, and lymph node), as has been shown in rodents. Glucocorticoids affect eosinophil infiltration into tissues by four mechanisms: (1) sequestration into lymphoid tissues, (2) induction of eosinophil apoptosis, (3) reduction of eosinophil production by bone marrow, and (4) alterations in the production of the cytokines/chemokines important in eosinophil trafficking.158–160 Glucocorticoids suppress the production of several cytokines important for the induction of adhesion molecules on endothelial cells, including IL-1, TNF-α, IL-4, and IL-13, and the release of eosinophil-active chemokines, including CCL5, CCL7, and CCL11. Unfortunately, “steroid resistance” develops in some patients, and long-term administration of glucocorticoids is associated with limiting side effects. The mechanism of glucocorticoid resistance is unclear but, in part, may be explained by decreased numbers of glucocorticoid receptors, and polymorphisms and alterations in transcription factor activator protein-1 (AP-1).161 Patients who have or who develop glucocorticoid resistance require other therapies. Interestingly, lidocaine and sulfonylureas (such as glyburide) also inhibit cytokine-induced eosinophil survival with glucocorticomimetic effects and may have antieosinophilic effects clinically.162–164

Calcineurin antagonists, such as cyclosporine, tacrolimus, and pimecrolimus, broadly inhibit T-cell cytokine release including those that specifically induce eosinophilic inflammation (IL-4, IL-5, and GM-CSF). They also decrease the expression of CCL5, CCL11, and IL-5 with associated decreased tissue eosinophilia as has been shown in atopic dermatitis.165 Mammalian target of rapamycin (mTOR) inhibitors, including rapamycin, have direct effects on eosinophils, decreasing eosinophil granule protein release after IL-5 activation.166 The use of these therapeutic agents are limited by their side effects including immunosuppression, as well as other metabolic effects that may be in part genetically determined.167

Several myelosuppressive drugs, including hydroxyurea, vincristine sulfate, cyclophosphamide, methotrexate, 6-thioguanine, 2-chlorodeoxyadenosine and cytarabine combination therapy, pulsed chlorambucil, and etoposide, may be beneficial in eosinophil-associated disease alone or as steroid-sparing agents. Hydroxyurea has been particularly effective in decreasing circulating eosinophil numbers.

In myeloproliferative hypereosinophilic syndrome (chronic eosinophilic leukemia) with the FIP1L1-PDGFRA mutation that codes for a tyrosine kinase, imatinib mesylate, a tyrosine kinase inhibitor, is approved for the treatment of chronic myelogenous leukemia and the hypereosinophilic syndrome and has produced rapid, complete or near complete remissions.168 Patients who have features of myeloproliferative HES but who lack FIP1L1-PDGFRA still may respond to imatinib (see Chapter 36).169

Alemtuzumab is a monoclonal antibody to CD52 that is used to deplete CD52+ lymphocytes in the treatment of chronic (B-cell) lymphocytic leukemia and T-cell lymphoma. Eosinophils, but not neutrophils, also express CD52, and alemtuzumab has been useful in treating patients with refractory hypereosinophilic syndrome, including those with abnormal T cells,170–172 but has serious limiting side effects from cytopenias, infusion reactions and infections.

Interferons, both IFN-α and IFN-γ, may be therapeutically beneficial in eosinophil-associated disease by inhibiting eosinophil degranulation and inflammatory mediator release. IFN-α may be better tolerated than IFN-γ and is used as a steroid-sparing agent predominantly in patients with lymphocytic variant hypereosinophilic syndrome, but also may be useful in myeloproliferative variant hypereosinophilic syndrome (see Chapter 36).173,174

Other Therapeutic Agents

PRÉCIS

Acute febrile neutrophilic dermatosis was originally described by Dr. Robert Douglas Sweet in the August–September 1964 issue of the British Journal of Dermatology. The cardinal features of “a distinctive and fairly severe illness” that had been encountered in eight women during the 15-year period from 1949 to 1964 were summarized. Although the condition was originally known as the Gomm–Button disease “in eponymous honor of the first two patients” with the disease in Dr. Sweet's department, “Sweet's syndrome” has become the established eponym for this acute febrile neutrophilic dermatosis.1–10

More than 1,000 cases of Sweet syndrome have been reported since Sweet's original paper.1–509 The distribution of Sweet syndrome cases is worldwide and there is no racial predilection.1,2,12,16–20,30,31 The dermatosis presents in three clinical settings.13,15

Diagnostic criteria for classical or idiopathic Sweet syndrome were proposed by Su and Liu in 1986 and modified by von den Driesch in 1994 (Table 32-1).11–14 It may be associated with infection (upper respiratory tract or gastrointestinal tract), inflammatory bowel disease, or pregnancy.13,15 Two studies have noted a seasonal preference for the onset of Sweet syndrome for either autumn or spring in 70% of 42 patients.416 or autumn.496

Classical Sweet syndrome most commonly occurs in women between the ages of 30 to 60 years. However, classical Sweet syndrome also occurs in younger adults and children.32–48,405,415,445,447,453 The youngest Sweet syndrome patients are brothers who developed the dermatosis at 10 and 15 days of age.46

Several investigators consider it appropriate to distinguish between the classical form and the malignancy-associated form of this disease since the onset or recurrence of many of the cases of Sweet syndrome are temporally associated with the discovery or relapse of cancer.15,49–60 Recently, the investigators of a comprehensive review of 66 pediatric Sweet syndrome patients observed that 44% of 30 children between 3 and 18 years of age had an associated hematologic malignancy.405,447 Malignancy-associated Sweet syndrome in adults does not have a female predominance and is most often associated with acute myelogenous leukemia.61,62 In Sweet syndrome patients with dermatosis-related solid tumors, carcinomas of the genitourinary organs, breast, and gastrointestinal tract are the most frequently occurring cancers.1,2,63–66

Criteria for drug-induced Sweet syndrome were established by Walker and Cohen in 1996 (Table 32-1).13 This variant of the dermatosis is most frequently observed to occur in association with the administration of granulocyte-colony stimulating factor (G-CSF).1,2,13,67,68 However, several other medications have also been implicated in eliciting drug-induced Sweet syndrome (^ch32etb1.1).11,13,17,39,41,69–124,401,402,422,427–429,436,437,439,446,455,456,463,464,468,469471,474,476,477,482,487, 489,490,492,493,505,506

The pathogenesis of Sweet syndrome may be multifactorial and remains to be definitively determined. A condition, similar to Sweet syndrome, presenting as a sterile neutrophilic dermatosis, has been described in a female standard poodle dog after treatment with the nonsteroidal anti-inflammatory drug firocoxib and in multiple dogs temporally associated with the administration of carprofen.403

Sweet syndrome may result from a hypersensitivity reaction to an eliciting bacterial, viral, or tumor antigen.2,127 A septic process is suggested by the accompanying fever and peripheral leukocytosis. Indeed, a febrile upper respiratory tract bacterial infection or tonsillitis may precede skin lesions by 1–3 weeks in patients with classic Sweet syndrome. Also, patients with Yersinia enterolitica intestinal infection-associated Sweet syndrome have improved with systemic antibiotics.2,77,125–127

The systemic manifestations of Sweet syndrome resemble those of familial Mediterranean fever. Recently, the simultaneous occurrence of both conditions has been observed.421 Also, in a patient with chronic myelogenous leukemia-associated Sweet syndrome, the causative gene mutation for familial Mediterranean fever was detected.448 Hence, the pathogenesis for these conditions may be similar.

Leukotactic mechanisms, dermal dendrocytes, circulating autoantibodies, immune complexes, human leukocyte antigen (HLA) serotypes, and cytokines have all been postulated to contribute to the pathogenesis of Sweet syndrome. Complement does not appear to be essential to the disease process. In some patients antibodies to neutrophilic cytoplasmic antigens (ANCAS) have been demonstrated;430 however, these are likely to represent an epiphenomenon.2

Cytokines—directly and/or indirectly—may have an etiologic role in the development of Sweet syndrome symptoms and lesions.2,21–23 Elevated serum levels of granulocyte-colony stimulating factor and interleukin-6 were detected in a patient with myelodysplastic syndrome-associated Sweet syndrome who was not receiving a drug.128 Detectable levels of intra-articular synovial fluid granulocyte macrophage-colony stimulating factor has also been observed in an infant with classical Sweet syndrome.44 Another study demonstrated that the serum G-CSF level was significantly higher in individuals with active Sweet syndrome than in dermatosis patients with inactive Sweet syndrome.129 And, a recent study showed that the level of endogenous G-CSF was closely associated with Sweet syndrome disease activity in a patient with acute myelogenous leukemia-associated Sweet syndrome and neutrophilic panniculitis.461

Significantly elevated levels of helper T-cell type 1 cytokines (interleukin-2 and interferon-γ) and normal levels of a helper T-cell type 2 cytokine (interleukin-4) have been seen in the sera of Sweet syndrome patients.130 In a patient with neuro-Sweet disease presenting with recurrent encephalomeningitis, serial measurements of cerebral spinal fluid interleukin-6, interferon-γ, interleukin-8, and IP10 [which is also referred to as the chemokine (C–X–C motif) ligand 10 (CXCL10)] were elevated as compared to levels in control subjects with neurologic disorders and also correlated with total cerebral spinal fluid cell counts; this data suggests an important role of the helper T-cell type 1 cell (whose cytokines include interferon-γ and IP10) and interleukin-8 (a specific neutrophil chemoattractant) in the pathogenesis of neuro-Sweet disease.478 Other studies showed decreased epidermal staining for interleukin-1 and interleukin-6 and postulated that this was due to the release of these cytokines into the dermis.131 In summary, G-CSF, granulocyte macrophage colony stimulating factor, interferon-γ, interleukin-1, interleukin-3, interleukin-6, and interleukin-8 are potential cytokine candidates in the pathogenesis of Sweet syndrome.2,13,21–23,44,128–132

History

Sweet syndrome patients may appear dramatically ill. The skin eruption is usually accompanied by fever and leukocytosis. However, the skin disease can follow the fever by several days to weeks or be concurrently present with the fever for the entire episode of the dermatosis. Arthralgia, general malaise, headache, and myalgia are other Sweet syndrome associated symptoms (Table 32-2).1,2,23

Cutaneous Lesions

Skin lesions of Sweet syndrome typically appear as tender, red or purple–red, papules or nodules. The eruption may present as a single lesion or multiple lesions that are often distributed asymmetrically (Fig. 32-1). The pronounced edema in the upper dermis of the lesions results in their transparent, vesicle-like appearance and has been described as an “illusion of vesiculation” (Fig. 32-2). In later stages, central clearing may lead to annular or arcuate patterns. The lesions may appear bullous, become ulcerated, and/or mimic the morphologic features of pyoderma gangrenosum in patients with malignancy-associated Sweet syndrome.133,134 The lesions enlarge over a period of days to weeks. Subsequently, they may coalesce and form irregular sharply bordered plaques (Fig. 32-3). They usually resolve, spontaneously or after treatment, without scarring. Lesions associated with recurrent episodes of Sweet syndrome occur in one-third to two-thirds of patients.1,2,135,136

Cutaneous pathergy, also referred to as skin hypersensitivity, is a dermatosis-associated feature.1,2 It occurs when Sweet syndrome skin lesions appear at sites of cutaneous trauma.458,496 These include the locations where procedures have been performed such as biopsies,20 injection sites,431 intravenous catheter placement,20 and venipuncture.12,17,20,37,137,138 They also include sites of insect bites and cat scratches,20 areas that have received radiation therapy,139–141,138,484 and places that have been contacted by sensitizing antigens.137,142,420 In addition, in some Sweet syndrome patients, lesions have been photodistributed or localized to the site of a prior phototoxic reaction (sunburn).13,20,98,143–145 Sweet syndrome lesions have also rarely been located on the arm affected by postmastectomy lymphedema.100,146,419,505

Sweet syndrome can present as a pustular dermatosis.147 The lesions appear as tiny pustules on the tops of the red papules or eythematous-based pustules. Some of the patients previously described as having the “pustular eruption of ulcerative colitis” are perhaps more appropriately included in this clinical variant of Sweet syndrome.1,148

“Neutrophilic dermatosis of the dorsal hands” or “pustular vasculitis of the dorsal hands” refers to a localized, pustular variant of Sweet syndrome when the clinical lesions are predominantly restricted to the dorsal aspect of the hands.3,149–154 The lesions from this latter group of individuals are similar to those of Sweet syndrome in morphology and rapid resolution after systemic corticosteroids and/or dapsone therapy was initiated. In addition, many of the individuals with this form of the disease also had concurrent lesions that were located on their oral mucosa, arm, leg, back, and/or face.3,155–163,425,435,440,442,457,462,470,495,497

The cutaneous lesions of subcutaneous Sweet syndrome usually present as erythematous, tender dermal nodules on the extremities.4,8,12,17,99,119,164–185 When the lesions are located on the legs, they often mimic erythema nodosum.170 Since Sweet syndrome can present concurrently21,125,187–189 or sequentially170 with erythema nodosum,17,21,187,190,509 tissue evaluation of one or more new dermal nodules may be necessary to establish the correct diagnosis—even in a patient whose Sweet syndrome has previously been biopsy-confirmed.1,2,4

Related Physical Findings

Extracutaneous Manifestations

(^ch32etb2.1.) Extracutaneous manifestations of Sweet syndrome may include the bones, central nervous system, ears, eyes, kidneys, intestines, liver, heart, lung, mouth, muscles, and spleen.12,16,17,20,25,26,32,33,44,73,75,101,117,138,139,165,202,203,205,212–257,407,408,410,433,444,450,452,459,465,468,475,478,481,486,509 The incidence of ocular involvement (such as conjunctivitis) is variable in classical Sweet syndrome and uncommon in the malignancy-associated and drug-induced forms of the dermatosis; however, it may be the presenting feature of the condition. Mucosal ulcers of the mouth occur more frequently in Sweet syndrome patients with hematologic disorders and are uncommon in patients with classical Sweet syndrome23,26,102,117,203,252; similar to extracutaneous manifestations of Sweet syndrome occurring at other sites, the oral lesions typically resolve after initiation of treatment with systemic corticosteroids.1,2 In children, dermatosis-related sterile osteomyelitis has been reported.

Associated Diseases

(eTable 32-2.2.) Several conditions have been observed to occur either before, concurrent with, or following the diagnosis of Sweet syndrome. Therefore, the development of Sweet syndrome may be etiologically related to Behcet's disease, cancer, erythema nodosum, infections, inflammatory bowel disease, pregnancy, relapsing polychondritis, rheumatoid arthritis, sarcoidosis, and thyroid disease. The association between Sweet syndrome and the other conditions (eTable 32-2.2) remains to be established.1,2,5,11–20,30,36–43,69–126, 158–161,164,166,186–190,195,214,231,236,259–339, 400,402,406,410,411,415,417,418,421–424,426–429,434,436–442,445,446,448,449,452–456,459, 460,463,464,466–469,471–477,479,480,482,483,487–490, 492–494,496,497,500,502–504,508,509

Associated Neutrophilic Dermatoses

An inflammatory infiltrate of mature polymorphonuclear leukocytes is the unifying characteristic of neutrophilic dermatoses of the skin and mucosa. Concurrent or sequential occurrence of Sweet syndrome with either erythema elevatum diutinum,340 neutrophilic eccrine hidradenitis,6 pyoderma gangrenosum,9,231,269,341,342,430,483,497 subcorneal pustular dermatosis,6,9 and/or vasculitis3,192,231 has been observed. Although these conditions can display similar clinical and pathologic features, the location of the neutrophilic infiltrate helps to differentiate them.6,120,499

Concurrent Leukemia Cutis

In patients with hematologic disorders, Sweet syndrome may present as a paraneoplastic syndrome (signaling the initial discovery of an unsuspected malignancy), a drug-induced dermatosis (following treatment with either all-trans-retinoic acid, bortezomib, G-CSF, or imatinib mesylate), or a condition whose skin lesions concurrently demonstrate leukemia cutis.1 Acute leukemia (myelocytic and promyelocytic) is the most frequent hematologic dyscrasia associated with leukemia cutis (characterized by abnormal neutrophils) and Sweet syndrome (consisting of mature polymorphonuclear leukocytes) being present in the same skin lesion.1,70,71,93,109,165,205–211,497 Myelodysplastic syndrome and myelogenous leukemia (either chronic or not otherwise specified) are the other associated hematologic disorders that have been associated with concurrent Sweet syndrome and leukemia cutis.109

“Secondary” leukemia cutis, in which the circulating immature myeloid precursor cells are innocent bystanders that have been recruited to the skin as the result of an inflammatory oncotactic phenomenon stimulated by the Sweet syndrome lesions has been suggested as one of the hypotheses to explain concurrent Sweet syndrome and leukemia cutis in the same lesion.165,206,207 Alternatively, “primary” leukemia cutis, in which the leukemic cells within the skin constitutes the bonified incipient presence of a specific leukemic infiltrate is another possibility.207 Finally, it is possible that the atypical cells of leukemia cutis developed into mature neutrophils of Sweet syndrome as a result of G-CSF therapy-induced differentiation of the sequestered leukemia cells in patients with “primary” leukemia cutis who were being treated with this agent.205

Histopathology

Evaluation of a lesional skin biopsy is helpful when the diagnosis of Sweet syndrome is suspected. Lesional tissue should also be submitted for bacterial, fungal, mycobacterial, and possibly viral cultures since the pathologic findings of Sweet syndrome are similar to those observed in cutaneous lesions caused by infectious agents.1,2

A diffuse infiltrate of mature neutrophils is characteristically present in the papillary and upper reticular dermis (Fig. 32-4); however, it can also involve the epidermis or adipose tissue. “Histiocytoid” Sweet syndrome refers to the setting in which the hematoxylin and eosin-stained infiltrate of immature myeloid cells are “histiocytoid-appearing” and are therefore initially misinterpreted as histiocytes.201,412–414,436,443,445,460,474

The dermal inflammation is usually dense and diffuse; however, it can also be perivascular or demonstrate “secondary” changes of leukocytoclastic vasculitis believed to be occurring as an epiphenomenon and not representative of a “primary” vasculitis,3,192,193 Neutrophilic spongiotic vesicles194 or subcorneal pustules12,80,167,195,196 result from exocytosis of neutrophils into the epidermis.12,17,80,167,194,195,197 When the neutrophils are located either entirely or only partially in the subcutaneous fat, the condition is referred to as “subcutaneous Sweet syndrome”.4,8,12,17,99,119,164–184,451,461,493,497

Edema in the dermis, swollen endothelial cells, dilated small blood vessels, and fragmented neutrophil nuclei (referred to as karyorrhexis or leukocytoclasia) may also be present (Fig. 32-5). Fibrin deposition or neutrophils within the vessel walls (changes of “primary” leukocytoclastic vasculitis) are usually absent and the overlying epidermis is normal.1,2,23,167,168 However, the spectrum of pathologic changes described in cutaneous lesions of Sweet syndrome has expanded to include concurrent leukemia cutis, vasculitis, and variability of the composition or the location of the inflammatory infiltrate.3,191,491,496

Lymphocytes or histiocytes may be present in the inflammatory infiltrate of Sweet syndrome lesions.11,104,167,168,198–200,504 Eosinophils have also been noted in the cutaneous lesions from some patients with either idiopathic11,167,168,195,202–204,212 or drug-induced84,107,110,111 Sweet syndrome. Abnormal neutrophils (leukemia cutis)—in addition to mature neutrophils—comprise the dermal infiltrate in occasional Sweet syndrome patients with hematologic disorders.1,70,71,93,109,165,205–211

Pathologic findings of Sweet syndrome can also occur in extracutaneous sites. Often, these present as sterile neutrophilic inflammation in the involved organ. These changes have been described in the bones, intestines, liver, aorta, lungs, and muscles of patients with Sweet syndrome.2

Other Laboratory Tests

Peripheral leukocytosis with neutrophilia and an elevated erythrocyte sedimentation rate and are the most consistent laboratory findings in Sweet syndrome.23 However, leukocytosis is not always present in patients with biopsy-confirmed Sweet syndrome.26 For example, anemia, neutropenia, and/or abnormal platelet counts may be observed in some of the patients with malignancy-associated Sweet syndrome. Therefore, a complete blood cell count with leukocyte differential and platelet count, evaluation of acute phase reactants (such as the erythrocyte sedimentation rate or C-reactive protein), serum chemistries (evaluating hepatic function and renal function), and a urinalysis should be performed. It is also reasonable to perform a serologic evaluation of thyroid function since there appears to be a strong association between thyroid disease and Sweet syndrome.1,2

Special Tests

Evaluation for Extracutaneous Manifestations

Extracutaneous manifestations of Sweet syndrome may result in other laboratory abnormalities. Patients with central nervous system involvement may have abnormalities on brain SPECTs (single photon emission computed tomography), computerized axial tomography, electroencephalograms, magnetic resonance imaging, and cerebrospinal fluid analysis. Patients with kidney and liver involvement may demonstrate urinalysis abnormalities (hematuria and proteinuria) and hepatic serum enzyme elevation. And, patients with pulmonary involvement may have pleural effusions and corticosteroid-responsive culture-negative infiltrates on their chest roentgenograms.2,343

Malignancy Workup

Recommendations for the initial malignancy workup in newly diagnosed Sweet syndrome patients without a prior cancer were proposed by Cohen and Kurzrock in 1993.15 Their recommendations were based upon the age-related recommendations of the American Cancer Society for early detection of cancer in asymptomatic persons and the neoplasms that had concurrently been present or subsequently developed in previously cancer-free Sweet syndrome patients. The recommended evaluation included the following:

1. A detailed medical history

2. A complete physical examination, including:

   1. examination of the thyroid, lymph nodes, oral cavity, and skin;

   2. digital rectal examination;

   3. breast, ovary, and pelvic examination in women; and

   4. prostate and testicle examination in men.

3. Laboratory evaluation:

   1. carcinoembryonic antigen level;

   2. complete blood cell count with leukocyte differential and platelet count;

   3. pap test in women;

   4. serum chemistries;

   5. stool guaiac slide test;

   6. urinalysis; and

   7. urine culture.

4. Other screening tests:

   1. chest roentgenograms;

   2. endometrial tissue sampling in either menopausal women or women with a history of abnormal uterine bleeding, estrogen therapy, failure to ovulate, infertility, or obesity; and

   3. sigmoidoscopy in patients over 50 years of age.

Since the initial appearance of dermatosis-related skin lesions had been reported to precede the diagnosis of a Sweet syndrome-associated hematologic malignancy by as long as 11 years, they also suggested that it was reasonable to check a complete blood cell count with leukocyte differential and platelet count every 6–12 months.2,15

Clinical Differential Diagnosis

Sweet syndrome skin and mucosal lesions mimic those of other conditions (Table 32-3.)2,15,23,148,165,202,220,344,345,409,421,448,498 Therefore, infectious and inflammatory disorders, neoplastic conditions, reactive erythemas, vasculitis, other cutaneous conditions, and other systemic diseases are included in the clinical differential diagnosis of Sweet syndrome.

Histologic Differential Diagnosis

The histologic differential diagnosis of Sweet syndrome includes conditions microscopically characterized by either neutrophilic dermatosis or neutrophilic panniculitis (^ch32etb3.1).2–4,6,12,193,346–353,432,451 The pathologic changes associated with Sweet syndrome are similar to those observed in an abscess or cellulitis; therefore, culture of lesional tissue for bacteria, fungi, and mycobacteria should be considered to rule out infection.23 Leukemia cutis not only mimics the dermal changes of Sweet syndrome, but can potentially occur within the same skin lesion as Sweet syndrome; however, in contrast to the mature polymorphonuclear neutrophils found in Sweet syndrome, the dermal infiltrate in leukemia cutis consists of malignant immature leukocytes.354 The pathologic changes in the adipose tissue of subcutaneous Sweet syndrome lesions can be found in either the lobules, the septae, or both; therefore, conditions characterized by a neutrophilic lobular panniculitis also need to always be considered and ruled out.2,4

Complications in patients with Sweet syndrome can be directly related to the mucocutaneous lesions or indirectly related to the Sweet syndrome-associated conditions. Skin lesions may become secondarily infected and antimicrobial therapy may be necessary. In patients with malignancy-associated Sweet syndrome, reappearance of the dermatosis may herald the unsuspected discovery that the cancer has recurred. Systemic manifestations of Sweet syndrome-related conditions—such as inflammatory bowel disease, sarcoidosis and thyroid diseases—may warrant disease-specific treatment.

The symptoms and lesions of Sweet syndrome eventually resolved without any therapeutic intervention in some patients with classical Sweet syndrome. However, the lesions may persist for weeks to months.10,23,254,355 In patients with malignancy-associated Sweet syndrome, successful management of the cancer occasionally results in clearing of the related dermatosis.13,15,23 Similarly, discontinuation of the associated medication in patients with drug-induced Sweet syndrome is typically followed by spontaneous improvement and subsequent resolution of the syndrome.13,15,23 Surgical intervention has also resulted in the resolution of Sweet syndrome in some of the patients who had associated tonsillitis, solid tumors, or renal failure.1,2,19,315,356,357,507

Sweet syndrome may recur following either spontaneous remission or therapy-induced clinical resolution.10 The duration of remission between recurrent episodes of the dermatosis is variable. Sweet syndrome recurrences are more common in cancer patients; in this patient population, the reappearance of dermatosis-associated symptoms and lesions may represent a paraneoplastic syndrome that is signaling the return of the previously treated malignancy.1,2,15,135

Systemic corticosteroids are the therapeutic mainstay for Sweet syndrome (^ch32etb3.2).7,8–10,12,16,17,19,20,23,36,49,50,70,184,223,233,240,250,284,358–361 Initiation of therapy promptly results in improvement of the symptoms and resolution of the mucocutaneous lesions. Daily pulse methylprednisolone administered intravenously may be necessary in patients with refractory disease. Topical (such as 0.05% clobetasol propionate)13,19–21,30,234,362–365 or intralesional (such as triamcinolone acetonide at a dose between 3.0 and 10.0 mg/cc)362,366,367 corticosteroids may be effective for treating localized Sweet syndrome lesions.1,2,7,9