Eosinophilia, Eosinophilic Neoplasms, and the Hypereosinophilic Syndromes

Eosinophilopoiesis

Eosinophils originate from pluripotent and granulocyte-committed, CD34 ⁺ , hematopoietic progenitor cells that are detectable in the BM and in the PB, and under certain conditions (local inflammation) also in various extramedullary organ sites. Eosinophil development from their multipotent and lineage-restricted progenitors is controlled by a network of transcription factors, including GATA-1, GATA-2, C/EBP-A, and C/EBP-E (see Chapter 9 ). Lineage-specific signaling as well as transcription factor expression is controlled by a network of cytokines and cytokine receptors. In fact, eosinophils develop from their progenitors in response to T-cell–derived growth factors, including IL-3, GM-CSF, and IL-5 (see Chapter 11 ). The most potent and specific eosinophilopoietic growth factor is IL-5. Eosinophils are derived from multipotent or bipotent colony-forming progenitor cell units (CFUs). Bipotent CFUs giving rise to eosinophils and basophils (CFU-eo/ba) are frequently detected in the BM and PB in healthy individuals, as well as in various disease states. By contrast, progenitor cells selectively giving rise to eosinophils and mast cells (CFU-eo/mast) are very rare. At an early stage of progenitor cell development, eosinophil-committed progenitors exhibit CD34 and CD38, and coexpress receptors for IL-3, GM-CSF, and IL-5. Later, CD34 expression is diminished, whereas the interleukin receptors continue to be expressed, and several additional, functionally important, cell-surface antigens are acquired by maturing eosinophils.

IL-3, GM-CSF, and IL-5 promote eosinophil differentiation of hematopoietic progenitor cells and eosinophil maturation of lineage-restricted precursors, as well as the recruitment, priming, activation, and survival of mature eosinophils. Whereas IL-3 and GM-CSF also affect other hematopoietic lineages, including basophils, IL-5 acts as a more lineage-restricted (eosinophil-targeting) cytokine. Therefore, IL-5 and IL-5 receptors have been recognized as major targets of therapy in diseases associated with (reactive) HE. The high-affinity receptor for IL-5 is composed of an IL-5–specific alpha chain (CD125) and a beta chain (CD131) that mediates intracellular signaling and is shared also by the high-affinity receptors for GM-CSF (CD116/CD131) and IL-3 (CD123/CD131; see Table 74.5 ) (see Chapter 11 ). Expression of high-affinity IL-5 receptors is an important prerequisite for specific cellular responses to “physiologic doses” of IL-5. Overexpression of IL-5 is observed in many reactive conditions and disease states associated with (reactive) HE. Correspondingly, IL-5 transgenic mice develop profound eosinophilia, and IL-5 knockout mice do not develop HE in response to allergic stimuli or a helminth infection. Under various reactive conditions accompanied by HE, IL-5 is produced primarily by activated Th2-type helper cells. However, although activated T-cells are considered a primary and most important source of eosinophil-triggering cytokines, like IL-5, other cell types, such as mast cells, macrophages, natural killer (NK) cells, and stromal cells, also produce these cytokines. In myeloid neoplasms with HE, clonal expansion of eosinophils usually is a growth factor-independent process. Moreover, depending on the underlying disease, neoplastic eosinophils may produce eosinophilopoietic cytokines in an autocrine manner, thereby augmenting tissue and blood HE.

Mobilization and Migration of Eosinophils to Sites of Inflammation

In healthy individuals as well as in reactive states, eosinophils circulate only briefly in the PB, and then transit to extravascular sites. Depending on environmental factors and demand, eosinophils localize preferentially in tissues and organs exposed to external environments, principally the submucosal membranes and loose connective tissues of the skin, GI tract, genital tract, and lungs. Depending on the underlying pathology, eosinophils can migrate into any vascularized organ. During an acute or chronic inflammatory process, larger numbers of eosinophils may be recruited actively into local tissue sites. The mobilization of eosinophils from the vasculature into (vascularized) tissues is a multistep process that involves their rolling on vascular endothelial cells and their subsequent adhesion to endothelium via L-selectin, followed by binding to intercellular adhesion molecule-1 (ICAM-1 = CD54) through a CD18/CD11a/b-dependent mechanism ( Chapter 14, Chapter 16 ). Expression of selectins on endothelial cells is triggered by histamine and other mediators, whereas expression of ICAM-1 on endothelial cells is induced by various pro-inflammatory cytokines, such as tumor necrosis factor-α (TNF-α). As both histamine and TNF-α are produced and provided by (activated) mast cells, these cells are thought to play a crucial role in eosinophil recruitment in various allergic and other inflammatory reactions. However, the recruitment of eosinophils is a complex process involving many different mechanisms and molecules, such as E-selectin, endothelial-leukocyte adhesion molecule 1 (ELAM-1 = CD62E) or vascular cell-adhesion molecule-1 (VCAM-1 = CD106). Recruitment of eosinophils is also mediated by adhesion to VCAM-1 via the β1 integrin very late activation antigen-4 (VLA-4 = CD49d/CD29), which is expressed by eosinophils but not by blood neutrophils.

Eosinophil migration is regulated by several different chemotactic factors, including various chemokines (SDF-1, eotaxins, IL-8, and others), the complement fragment C5a, PAF, and several eosinotropic cytokines, such as IL-5 or GM-CSF, which also prime eosinophils for enhanced migratory responses to PAF and IL-8. In addition, IL-2, leukotriene B ₄ , and the lymphocyte chemoattractant factor (LCF = IL-16) can induce eosinophil migration. At least for neoplastic eosinophils, a most potent chemoattractant appears to be SDF-1, a stromal cell-derived chemokine that binds to CXCR4 (CD184) on normal and neoplastic eosinophils. Other potent chemoattractants for (neoplastic) eosinophils are the eotaxins (eotaxin-1, -2, -3) that bind to and signal through the chemokine receptor CCR3 (CD193; see Table 74.5 ). Depending on the underlying disease, the cell type(s) involved, and the presence of comorbidities (triggering, e.g., local or systemic inflammation) all these chemotactic factors may act together to trigger eosinophil recruitment and accumulation in the affected (often inflamed) tissue sites. Likewise, in several reactive conditions, IL-5 and eotaxin may cooperate locally to promote eosinophil migration and tissue HE.

The primary source of eosinophil chemotactic factors often remains unknown. In many disease states accompanied by HE, eosinophil chemotactic factors may be produced by microenvironmental cells, including fibroblasts, endothelial cells, T-cells, macrophages, and mast cells. Several of these chemokines, like SDF-1, are produced by stromal cells in a paracrine fashion (see Chapter 12 ). For example, in neoplastic eosinophils, the F/P mutant triggers expression and release of oncostatin M (OSM), which in turn induces fibroblast and endothelial cell proliferation but also the expression and release of SDF-1 from these cells, thereby augmenting the accumulation of additional eosinophils. In this regard, it is noteworthy that the SDF-1 receptor CXCR4 is not only expressed on mature eosinophils but also on hematopoietic stem cells and eosinophil-committed precursor cells. Depending on the type of underlying neoplasm, eosinophil chemotactic cytokines may also be produced in an autocrine manner.

Role of Mast Cells in Eosinophil Recruitment and Accumulation

In various inflammatory and allergic reactions, several different interactions between eosinophils and mast cells have been postulated (see Chapter 75 ). For example, mast cells provide the immediate stimuli upon activation to initiate allergic inflammation, resulting in eosinophil recruitment during the late phase following an allergen encounter. In addition, mast cells and eosinophils are often increased in numbers in allergic reactions, but also in non-allergic inflammatory conditions and even in neoplastic states. Both cell types express a similar profile of chemotactic receptors, such as CCR3, and both respond to eotaxins and RANTES (CCL5), leading to their recruitment into local tissue sites (see Chapter 12 ). Mast cells and eosinophils also interact with each other via cell–cell contact and through several different soluble mediators. Likewise, mast cell-derived heparin binds and stabilizes various chemokines, including the eotaxins. In addition, mast cells are a source of various cytokines involved in the regulation of eosinophil adhesion, migration, and function. As previously mentioned, mast cells express and release histamine and TNF-α, and thereby facilitate eosinophil recruitment. Mast cell proteases appear to have a dual action on eosinophil functions. Mast cell-derived chymase suppresses eosinophil apoptosis and induces the release of IL-6 and various chemokines (CXCL1, CXCL8, CCL2) from eosinophils. Beta-tryptase, on the other hand, can cleave eotaxin and RANTES, and thereby may limit eosinophil chemotaxis. Apart from these interactions, eosinophils and mast cells may sometimes also be derived from the same stem or progenitor cell compartment, and therefore may express the same driver mutant, which has been documented for cases of F/P + chronic eosinophilic leukemia (CEL) with accompanying mast cell expansion, as well as for patients with KIT D816V+ systemic mastocytosis (SM) with concomitant HE (see Chapter 75 ). Therefore, mast cells and eosinophils may often appear as a dual-target cell population, both in reactive and neoplastic states, which is a critical point when considering the development of specific therapies. For example, simultaneous downregulation of eosinophil and mast cell activation might be achieved by certain inhibitory surface molecules, such as CD300a or Siglec-8. Several other target receptors, such as Siglec-3 (CD33), are expressed on eosinophils, mast cells, and basophils. In patients with F/P + neoplasms, imatinib can induce apoptosis in both eosinophils and mast cells, as both cell types express the mutant platelet-derived growth factor receptor (PDGFR) target in this neoplasm.

Pathogenesis of Eosinophil-Associated End-Organ Damage in Patients With Hypereosinophilic Syndrome

Regardless of the underlying etiology (neoplastic or reactive), sustained HE may lead to typical end-organ damage defined as HES. Even in patients with idiopathic HE, such organ damage may occur and is then called idiopathic HES (HES _I ). However, not all patients presenting with HE develop organ damage even if followed over a longer time period. For example, patients with an underlying allergic disease usually fail to develop the organ damage seen in patients with typical HES. When neither an underlying disease nor HE-related organ damage is found, the diagnosis HE of undetermined significance (HE _US ) is established. In general, the risk of HES is lower in reactive forms of HE, especially when the underlying disease can be treated successfully. Correspondingly, IL-5 transgenic mice with massive HE do not develop significant end-organ damage, suggesting that other factors (in addition to IL-5) are likely necessary for the occurrence of an HE-related organ damage. In patients with an underlying MPN or other underlying myeloid neoplasm, the risk of HE-related organ damage (HES) is relatively high. A particularly high risk is found in patients with mutations in PDGFR genes. Likewise, most untreated patients with F/P + MPN-eo or CEL may develop HES with cardiac involvement over time. Additional risk factors may also contribute to the development of HES in these patients, including, among others, an additional prothrombophilic state, late detection of the disease, and late initiation of specific (anti-CEL) therapy. There may also be a certain genetic predisposition for early occurrence of tissue HE in patients with F/P + CEL ( F/P + MPN-eo). In particular, the severity (extent) of tissue HE seems to correlate with a polymorphic variant (SNP in the 5′-UTR) in the IL-5RA gene. However, it remains unclear whether this SNP predisposes to HES development or progression, and thus is relevant clinically. Today, the general recommendation is to initiate therapy with imatinib early in all F/P + patients, regardless of genetic, molecular or other factors, in order to prevent any HES occurrence and to avoid disease progression (to acute leukemia).

A number of different organ systems are affected in patients with HES. The most common manifestations are summarized in Table 74.7 . Although multiple organ systems are involved and the manifestation patterns are quite complex with varying courses and outcomes, some common pathogenetic factors have been described. One common feature is the mobilization of the tissue microenvironment by eosinophil-derived mediators and cytokines, which leads to tissue remodeling, fibrosis, and increased angiogenesis. Another common factor is tissue inflammation, which is typically triggered by eosinophil-derived cytokines and chemokines. Finally, HE-related organopathies may often result from direct cytotoxic damage that is induced by the various toxic mediators produced and released by (reactive and neoplastic) eosinophils. With regard to thromboembolic events, several different eosinophil-derived mediators have been implicated, including TNF-α, plasminogen activator inhibitors (PAIs), and eosinophil-derived DNA traps (see Table 74.1 ). In addition, ECPs have the capacity to alter thrombomodulin activity. A number of different compounds produced by eosinophils, including cationic proteins and various cytokines, may induce fibrosis. Among these cytokines, OSM may play a particular role as a HES-related profibrogenic cytokine. HE-related neoangiogenesis may be triggered by IL-8, OSM, and vascular endothelial growth factor (VEGF). All these cytokines are produced and secreted by normal/reactive and also by neoplastic eosinophils. However, neoplastic eosinophils produce some of these cytokines in excess over their normal counterparts. For example, F/P + eosinophils express and release huge amounts of OSM, IL-8, and VEGF. The F/P mutant induces the production and secretion of these mediators in neoplastic cells in a STAT5-dependent manner. Concerning the mobilization of the microenvironment in various organs in patients with HES, a most critical F/P target gene may be OSM . Notably, OSM initiates not only fibrosis and angiogenesis by triggering fibroblast proliferation and tube formation in endothelial cells, but also by inducing the production of multiple cytokines and chemokines in stromal cells. Among these mediators are IL-8, SDF-1, and other chemotactic factors that can recruit additional CXCR4+ precursor cells and eosinophils to local tissue sites.

An unresolved question is why HE-related fibrosis and thrombosis affect certain tissues and organs, including the heart and lungs. One explanation may be that eosinophils invade heart and lung tissues via “physiologic” routes, which may not be the case in other MPNs where HE usually does not occur. An additional explanation may be that eosinophil-derived mediators are toxic to certain (tissue-specific) cell types such as cardiomyocytes. Indeed, as noted earlier, eosinophils express a number of different cationic proteins capable of inducing endothelial and endocardial damage and neurotoxicity. In addition, the eosinophil has the capacity to generate reactive oxidative species that can augment tissue damage.

Definition of Hypereosinophilia

HE is currently defined as a persistent, microscopically confirmed increase in PB eosinophils to ≥ 1.5 × 10 ⁹ /L. In previous definitions, persistent eosinophilia over 6 months was required to establish the diagnosis HE. However, based on improved diagnostics and treatment, and the necessity to introduce such treatment rapidly to avoid organ damage, the proposed definition of HE has been revised to a 4-week observation interval by the ICOG-EO group. The term “tissue HE” has also been proposed by several experts, and it may be useful to apply it in certain conditions, especially in the context of (to demonstrate) HE-related organ damage (HES). However, isolated tissue HE (without blood HE) is rare, and the documentation of tissue HE often requires special immunostains directed against eosinophil granule proteins. Therefore, the presence of tissue HE is often overlooked or is not documented, and PB HE clearly remains the key diagnostic marker in daily practice. In some of these patients, criteria for blood HE are not met, but molecular and clinical signs are strongly indicative of a particular eosinophil disorder with or without accompanying HES. These patients should be followed closely, as they may progress to an overt eosinophil disease (neoplasm) over time. The same holds true for WHO-based neoplasms presenting with PDGFR or FGFR1 fusion genes without overt HE. In fact, many patients with FGFR1 -rearranged neoplasms do not have eosinophilia at presentation.

Diagnostic Algorithm

When arriving at the provisional pre-diagnostic “checkpoint” HE, two critical questions have to be addressed in order to approach to a final diagnosis: (1) is there any underlying disease or condition triggering HE? and (2) are there clinical signs and symptoms or laboratory abnormalities that point to the presence of HE-induced organ damage, so that the diagnosis of HES can be established? For example, the hematologic work-up in a patient with HE revealed CEL, and staging investigations showed endocardial thrombus formation: the final diagnosis in this patient is CEL with concomitant primary HES (HES _N ). A diagnostic algorithm is shown in Fig. 74.3 . In patients with typical HES and typical clinical manifestation, histopathologic evaluation is generally not warranted to confirm the presence of tissue HE. However, in patients with rare or atypical manifestations, such as renal failure, isolated myocarditis or bloody diarrhea, a tissue biopsy is required to document the presence of tissue HE and to confirm HE-induced organ damage, so that the diagnosis HES can be established. The demonstration of extensive extracellular deposition of eosinophil-derived proteins supports the conclusion the organ damage is “HE-related.”

Classification of Hypereosinophilia and Hypereosinophilic Syndrome

The classification of HE and HES is essentially based on the related disease and underlying etiology. Accordingly, HE and HES are both divided into a familial variant (HE _F , HES _F ), a primary/neoplastic form (HE _N , HES _N ), and reactive entities (HE _R , HES _R ; see Table 74.3 ). In a smaller fraction of patients, the etiology of HE and HES remains uncertain. In other rare cases, two different causative factors are detected. When the etiology of HE remains uncertain after a thorough investigation and appropriate testing and no organ damage is detected, the final diagnosis is HE of uncertain (undetermined) significance (HE _US ); and if HES is diagnosed, but no underlying disease is detected, the final diagnosis is idiopathic HES (HES _I ).

In general, four major groups of underlying disorders (conditions) have to be considered in HE (HES) patients: (1) myeloid (and rarely myeloid/lymphoid) neoplasms including eosinophilic leukemia (HE _N ; HES _N ), (2) lymphoid or nonhematopoietic neoplasms (paraneoplastic HE, a special variant of HE _R ), (3) common allergic, reactive, or immunologic conditions (HE _R and HES _R ), and (4) clinically defined, rare syndromes and conditions accompanied by HE, including rare inherited disorders and organ-restricted inflammatory diseases with tissue HE. Overall, the classification of HE/HES remains an important step in the diagnostic algorithm (see Fig. 74.3 ). Using generally accepted criteria and parameters, this approach should lead to a final diagnosis concerning the presence or absence of HES and the presence or absence of an underlying disease.