Eosinophilia: secondary, clonal and idiopathic


Ayalew Tefferi, Division of Hematology, Department of Internal Medicine, Mayo Clinic College of Medicine, 200 First Street SW, Rochester, MN 55905, USA. E-mail: tefferi.ayalew@mayo.edu


Blood eosinophilia signifies either a cytokine-mediated reactive phenomenon (secondary) or an integral phenotype of an underlying haematological neoplasm (primary). Secondary eosinophilia is usually associated with parasitosis in Third World countries and allergic conditions in the West. Primary eosinophilia is operationally classified as being clonal or idiopathic, depending on the respective presence or absence of a molecular, cytogenetic or histological evidence for a myeloid malignancy. The current communication features a comprehensive clinical summary of both secondary and primary eosinophilic disorders with emphasis on recent developments in molecular pathogenesis and treatment.

‘Theia yielded to Hyperion's love and gave birth to great Helios and bright Selene and Eos, who brings light to all mortals of this earth and to the immortal gods who rule the wide sky.’– Hesiod, Theogony, 371–74

Hyperion and Theia were Titans in Greek mythology who gave birth to three daughters: Helios, the goddess of the Sun; Selene, the goddess of the Moon; and Eos, the goddess of Dawn. Since time immemorial, Greek poets such as Homer romanticised with the colours of dawn, whose glorious reddish-crimson hues led to the naming of the first of the synthetic aniline dyes, discovered by W.H. Perkins (1838–1907) in 1856, as Eosin (Silverstein, 2005). Subsequently, the great Paul Ehrlich (1854–1915), who won the 1908 Nobel Prize for Physiology or Medicine, pioneered the use of chemical dyes as selective biological stains for the study of human tissue (Silverstein, 2005). Based on the specific affinities of certain blood cells for either basophilic or acidophilic dyes, Ehrlich defined and named several aniline-reactive leucocytes including the eosinophil (1879) and mast cells (1878) (Perkins, 1879; Crivellato et al, 2003). The selective action of aniline dyes on cells and tissues suggested to Ehrlich the possibility of creating ‘magic bullets’ that specifically target disease organisms while sparing normal tissue. It is ironic that this very concept has now been fully realised for platelet-derived growth factor receptor (PDGFR)-rearranged eosinophilic disorders, where treatment with Gleevec® (imatinib mesylate; STI571) produces complete molecular remission with minimal toxicity to normal tissue (Pardanani & Tefferi, 2004). Gleevec® is a 2-phenylaminopyrimidine tyrosine kinase inhibitor with specific activity for the Abelson tyrosine kinase (ABL), PDGFR and stem cell factor receptor (KIT).

Eosinophils are derived from haematopoietic stem cells that are committed initially to the myeloid and subsequently to the basophil–eosinophil lineage (Denburg et al, 1985). Cationic proteins (major basic protein, eosinophilic cationic protein, eosinophil-derived neurotoxin, eosinophil peroxidase), cytokines (interleukins and tumour necrosis factor) and lipid mediators (leucotrienes) constitute the content of the eosinophilic granule and mediate parasite defence reaction, allergic response, tissue inflammation and immune modulation (Rothenberg & Hogan, 2005). Interleukin (IL)-3, IL-5 and granulocyte monocyte-colony stimulating factor (GM-CSF) are considered to be the major eosinophil growth and survival factors and are coded by closely situated genes on chromosome 5q31–q33, a cytokine gene cluster that has also been linked to familial eosinophilia (Rioux et al, 1998). Both types 1 and 2 T-helper (Th1 and Th2) cells participate in early eosinophil development (GM-CSF and IL-3) whereas Th2-derived IL-5 appears to be crucial in robust eosinophil proliferation (Wierenga et al, 1993). On the other hand, eosinophil chemotaxis and tissue access are facilitated by C-C chemokines (eotaxin, CCR3, PAF and RANTES) and endothelial adhesion molecules (integrins and vascular cell adhesion molecules) (Garcia-Zepeda et al, 1996; Gurish et al, 2002; Pope et al, 2005). In addition, eotaxin has been shown to play a collaborative role, along with IL-3 and IL-5, in early eosinophilopoiesis from murine embryonic stem cells (Hamaguchi-Tsuru et al, 2004). Eotaxin and its receptor, CCR3, may also be involved in embryonic myelopoiesis, in general, and their differentiation into mast cells, in particular (Quackenbush et al, 1998).

In health, the upper limits for peripheral blood eosinophil percentage and absolute eosinophil count (AEC) do not exceed 5% and 0·5 × 109/l respectively (Brigden & Graydon, 1997). Increased levels signify underlying disease and the degree of eosinophilia is arbitrarily assigned as mild (AEC 0·5–1·5 × 109/l), moderate (1·5–5 × 109/l) or severe (>5 × 109/l) (Brito-Babapulle, 2003). Most instances of eosinophilia are acquired although familial eosinophilia, an autosomal dominant disorder that is characterised by a stable eosinophil count and a relatively benign clinical course, has rarely been described (Klion et al, 2004a). From the early 20th century onwards, the association of blood eosinophilia with parasitosis and allergic diseases, both of which are IL-5 driven (Korenaga et al, 1991), had been well recognised and the distinction between such secondary cases and idiopathic eosinophilia was formalised in 1968 when Hardy and Anderson first introduced the term hypereosinophilic syndrome (HES) (Hardy & Anderson, 1968). At the same time, the occurrence of sometimes marked eosinophilia in association with certain forms of leukaemia and myeloproliferative disorders (MPD) did not go unnoticed (Gray & Shaw, 1949; Spitzer & Garson, 1973; Ellman et al, 1979). Accordingly, acquired eosinophilia is currently classified into secondary (cytokine-driven reactive phenomenon), clonal (presence of a bone marrow histological, cytogenetic or molecular marker of a myeloid malignancy) and idiopathic (neither secondary nor clonal) categories (Table I). HES is a subset of idiopathic eosinophilia that requires the presence of an AEC of >1·5 × 109/l as well as evidence for target organ damage. In addition to these features, the World Health Organization (WHO) definition of HES requires the absence of aberrant cytokine-secreting T-cell population (Bain et al, 2001).

Table I.   Causes of acquired eosinophilia.
(1) Secondary
 (a) Infections (mostly helminthic)
 (b) Drugs (anticonvulsants, antibiotics, sulpha drugs, antirheumatics, allopurinol, food allergy)
 (c) The pulmonary eosinophilias (see Table IV)
 (d) Miscellaneous other causes of autoimmune/inflammatory/ toxic origin
  (i) Eosinophilia-myalgia syndrome, toxic oil syndrome
  (ii) Eosinophilic fasciitis (a.k.a. Schulman syndrome), Kimura disease, Wells syndrome, Omenn syndrome
  (iii) Connective tissue diseases (scleroderma, polyarteritis nodosa, etc.)
  (iv) Sarcoidosis, inflammatory bowel disease, chronic pancreatitis
 (e) Malignancy (metastatic cancer, Hodgkin lymphoma)
 (f) Endocrinopathies (Addison disease, growth factor deficiency, etc.)
(2) Clonal
 (a) Acute leukaemia (both myeloid and lymphoblastic)
 (b) Chronic myeloid disorder
  (i) Molecularly defined
   (1) Bcr/Abl+ chronic myeloid leukaemia
   (2) PDGFRA-rearranged eosinophilic disorder (SM-CEL)
   (3) PDGFRB-rearranged eosinophilic disorder
   (4) KIT-mutated systemic mastocytosis
   (5) 8p11 syndrome
  (ii) Clinicopathologically assigned
   (1) Myelodysplastic syndrome
   (2) Myeloproliferative disorder
    (a) Classic myeloproliferative disorder (polycythaemia vera, etc.)
    (b) Atypical myeloproliferative disorder
     (i) Chronic eosinophilic leukaemia
     (ii) Systemic mastocytosis
     (iii) Chronic myelomonocytic leukaemia
     (iv) Unclassified myeloproliferative disorder
(3) Idiopathic including hypereosinophilic syndrome

Secondary eosinophilia


The most frequent cause of secondary eosinophilia worldwide is tissue-invasive parasitosis that includes infections with roundworms (nematodes), tapeworms (cestodes) and flukes (trematodes). Table II features representative organisms in this regard with their clinical presentation, geographical distribution and treatment drugs of choice (Zinkham, 1978; Hussain et al, 1981; Milder et al, 1981; Maxwell et al, 1987; Evengard, 1990; Pozio et al, 1993; Arjona et al, 1995; Uchiyama et al, 1999). An elaborate travel history that includes specific geography and exposure history to animals, insects, raw food or untreated water is key for guiding pertinent laboratory testing for suspected parasitosis. Infections other than those caused by worms (helminths) are infrequently associated with eosinophilia and involve certain (toxoplasmosis, isosporiasis, Dientamoeba fragilis infection) but not other (Malaria, giardiasis, Entamoeba histolytica infection) protozoans, Borrelia burgdorferi (a spirochete bacterium) and human immunodeficiency virus (HIV) (Grant & Klein, 1987; Junod, 1988; Granter et al, 1996; Tietz et al, 1997; Windsor & Johnson, 1999). In general, parasites that are isolated in either the intestinal lumen (tapeworms, ascaris) or an intact cyst (Echinococcus granulosus, cysticercosis) do not cause blood eosinophilia unless they are systemically introduced through tissue invasion or cyst disruption (Windsor & Johnson, 1999; Leder & Weller, 2000).

Table II.   Parasitosis associated with eosinophilia.
Disease/agentClinical featuresGeographical distributionTreatment
Lymphatic filariasis (roundworm)Elephantiasis
Pulmonary tropical eosinophilia
Tropics, subtropics, AsiaDiethylcarbamazine
Loa loa filariasis (roundworm)Subconjunctival worms
Skin lesions, episodic angioedema
Western AfricaDiethylcarbamazine
Onchocerca filariasis (roundworm)Skin nodules
Africa, Latin AmericaIvermectin
Gnathostomiasis (roundworm)Cutaneous larva migrans
Visceral larva migrans (meningitis)
Asia, MexicoAlbendazole
Anisakiasis (roundworm)Acute abdominal painWorldwide
Raw fish ingestion
Endoscopic removal of larvae
Hookworm (roundworm)
Ancylostoma duodenale
Necator americanus
Iron deficiency anaemiaWorldwide
Africa, Asia, the Americas
Australia, Middle East
Pyrantel pamoate
Ascariasis (roundworm)Abdominal pain, oral expulsion
Intestinal or biliary obstruction
Loeffler syndrome (pneumonitis)
Tropics, Subtropics
Rural southeastern US
Pyrantel pamoate
Strongyloidiasis (roundworm)Frequently asymptomatic
Abdominal pain, diarrhoea
Loeffler syndrome (pneumonitis)
Tropics, Subtropics
Rural South US
Trichinosis (roundworm)Intestinal symptoms, myositis myocarditis, conjunctivitisWorldwide
Europe, US
Toxocariasis (roundworm)Visceral and ocular larva migransWorldwideAlbendazole
Angiostrongyliasis (roundworm)Eosinophilic meningitisSoutheast Asia
Pacific Basin
No effective treatment
Paragonimiasis (lung fluke)Pneumonia, rusty sputum
Haemoptysis, skin lesions
Far East, Asia, Latin America, AfricaPraziquantel
Fascioliasis (liver fluke)Hepatitis, hepatomegaly
Biliary obstruction
Raw watercress ingestion
HaematobiumHaematuria, bladder cancerAfrica, Middle EastPraziquantel
MansoniPortal hypertensionLatin America, CarribeanOxamniquine
JaponicumPortal hypertensionFar EastPraziquantel
Isosporiasis (coccidian parasite)Chronic diarrhoeaWorldwide immunocompromised hostsTrimethoprim- sulphamethoxazole

Repeated stool examinations are critical for the diagnosis of parasite infestations that have intestinal phases in their life cycles and should be done regardless of the presence or absence of focal findings (Leder & Weller, 2000). In addition to stool, ova and parasites are sought in duodenal aspirate (strongyloidiasis, ascariasis, ancylostomiasis, clonorchiasis, fascioliasis) and sputum (strongyloidiasis, ascariasis, paragonimiasis, ancylostomiasis, schistosomiasis) as clinically indicated. Other laboratory investigations for suspected parasitosis include serology for schistosomiasis, paragonimiasis, filariasis, strongyloidiasis and toxocariasis. Furthermore, the presence of focal findings warrants imaging studies, spinal fluid analysis, blood smear examinations (filariasis), urine test (schistosomiasis, filariasis) and tissue biopsy (muscle biopsy for trichinellosis, liver or bladder biopsy for schistosomiasis) (Leder & Weller, 2000).


Different manifestations of drug-induced eosinophilia are summarised in Table III and some, such as the DRESS syndrome (drug rash with eosinophilia and systemic symptoms) are potentially fatal (Choquet-Kastylevsky et al, 1998, 2001; Britschgi et al, 2001; Carroll et al, 2001; Baraniuk & Maibach, 2005, Ten et al, 1988; Roujeau, 2005). Drug-induced skin lesions might or might not be accompanied by fever and are markedly heterogeneous in their appearance; generalised rash, Stevens–Johnson syndrome or toxic epidermal necrolysis (Wolf et al, 2005). Differential diagnosis of a suspected drug reaction includes infection (viral, bacterial, fungal), neoplastic or paraneoplastic manifestation (e.g. lymphoma, leukaemia, Sweet syndrome) and autoimmune/inflammatory conditions (e.g. connective tissue disease, serum sickness, Kawasaki disease). The systemic symptoms and signs of DRESS include fever, extensive rash, lymphadenopathy, pneumonia, hepatitis, arthritis and renal dysfunction. Other characteristic features include delayed onset (2–6 weeks after the first drug use) and the presence of atypical lymphocytosis. DRESS-associated drugs include cephalosporins (Akcam et al, 2005), vancomycin (Zuliani et al, 2005), nevirapine (Bourezane et al, 1998), phenobarbital (Baruzzi et al, 2003), carbamazepine (Descamps et al, 2001), phenytoin (Allam et al, 2004), minocycline (Knowles et al, 1996), allopurinol (Markel, 2005), sulfasalazine (Michel et al, 2005) and dapsone (Itha et al, 2003). Reactivation of human herpes virus 6 (HHV6) has been linked to severe DRESS (Descamps et al, 2001).

Table III.   Drug-induced eosinophilic syndromes.
Generalised rash with or without feverAny drug is a possibility
Mostly seen with antibiotics
Interstitial nephritis with eosinophiluriaAntibiotics, gold compounds, allopurinol
Pulmonary infiltratesNitrofurantoin, minocycline, naproxen, penicillins, phenylbutazone, sulindac, piroxicam, sulphonamides, nimesulide, tolfenamic acid
Pleuropulmonary manifestationsDantrolene sodium, bleomycin, methotrexate
HepatitisPhenothiazines, penicillins, tolbutamide, allopurinol, methotrexate, fluoroquinolones
Leucocytoclastic vasculitisAllopurinol, phenytoin
Chronic rhinosinusitis with nasal polyposis and asthmaAspirin
Eosinophilia-myalgia syndromel-tryptophan
DRESS syndrome (drug rash with eosinophilia and systemic symptoms)Carbamazepine, allopurinol, antibiotics, etc.

The pulmonary eosinophilias

A few eosinophilic processes are characterised by pulmonary lesions that are histologically composed of eosinophilic infiltrates that might or might not be accompanied by vasculitis, granulomas and microorganisms including fungi (Table IV) (Johkoh et al, 2000; Alberts, 2004; Grossi et al, 2004; Kalomenidis & Light, 2004; Khoo & Lim, 2004; Norman et al, 2004; Shorr et al, 2004; Takafuji & Nakagawa, 2004; Bargagli et al, 2005; Cottin & Cordier, 2005; Magnaval & Berry, 2005). One example is allergic bronchopulmonary aspergillosis (ABPA), which is a complication of long-standing asthma or cystic fibrosis, where inhaled conidia from Aspergillus fumigatus induce a host immune reaction that consists of airway hyper-reactivity, pulmonary infiltrates with fluctuating shadows and proximal bronchiectasias (Glimp & Bayer, 1983; Kumar & Gaur, 2000; Soubani & Chandrasekar, 2002). Diagnosis is established by documenting an immediate skin reaction to aspergillus antigens as well as increased levels of A. fumigatus-specific IgG and IgE immunoglobulins and increased total serum IgE level. Treatment consists of systemic steroids and azoles (Salez et al, 1999). ABPA can sometimes progress into a necrotising pneumonia (bronchocentric granulomatosis) that can also occur in the absence of ABPA (Yousem, 1997).

Table IV.   The pulmonary eosinophilic syndromes.
DiagnosisPeripheral eosinophiliaRadiologyBAL/biopsy findingsSystemic features
  1. ARDS, acute respiratory distress syndrome; BAL, bronchoalveolar lavage; IEP, idiopathic eosinophilic pneumonia; ABPA, allergic bronchopulmonary aspergillosis; CNS, central nervous system.

Chronic IEPMarkedPeripheral opacities
Migratory infiltrates
Marked eosinophiliaNon-specific, cough
Weight loss
Acute IEPMild to absentBilateral infiltrates
Pleural effusions
Marked eosinophiliaARDS
Recent-onset smoking
Churg–Strauss syndromeMarkedNon-specific
Migratory infiltrates
Sometimes normal
Asthma, rhinosinusitis, peripheral neuropathy
Cardiac and renal disease
Palpable purpura
Hypereosinophilic syndromeMarkedInterstitial infiltrates
Pulmonary nodules
Pleural effusions
CNS vasculitis
Tropical pulmonary eosinophilia (microfilariae)MarkedBilateral opacitiesEosinophilsFever, cough, hyper-reactive airways
ABPAModerateMucus plugs
Centrilobular nodules
Proximal bronchiectasis
Fungal hyphae
Allergic mucin
Asthma, Rhinosinusitis
Cystic fibrosis
Drug inducedMildAlveolar infiltrates
Pleural effusions
EosinophilsFever, rash
Radiation inducedMild to moderateUnilateral infiltratesEosinophilsFever, cough, dyspnoea

Churg–Strauss Syndrome (CSS) is a systemic vasculitis that involves small and medium vessel arteries and is characteristically accompanied by asthma and blood eosinophilia (Abril et al, 2003). In addition, many patients manifest rhinosinusitis, nasal polyposis, mononeuritis multiplex and palpable purpura. Less frequently encountered complications included pericarditis, myocardial disease and renal failure. Treatment for CSS includes corticosteroids and other immunosuppressive drugs. Approximately 50% of CSS patients display circulating antibodies to neutrophil cytoplasmic enzymes. Differential diagnosis in this instance includes Wegener granulomatosis and microscopic polyangiitis. Other forms of pulmonary vasculitic syndromes include giant cell arteritis, Takayasu arteritis and those associated with connective tissue disorders, such as polyarteritis nodosa, scleroderma, systemic lupus erythematosus and polymyositis.

Certain pulmonary eosinophilic syndromes are operationally classified as being idiopathic and include simple pulmonary eosinophilia (a.k.a. Loffler pneumonia) (Loffler, 1932) and acute or chronic eosinophilic pneumonia (Table IV). Simple pulmonary eosinophilia is a self-limiting syndrome with fluctuating pulmonary infiltrates and blood eosinophilia. Most such cases are currently linked to drug reactions or parasite infections. Acute idiopathic eosinophilic pneumonia (acute IEP) is a rare disorder that presents with a corticosteroid-responsive acute respiratory distress syndrome (ARDS) that is associated with marked bronchoalveolar lavage (BAL) eosinophilia and might represent an unusual host reaction to recent-onset smoking, drugs, infections or inhaled toxins or dust (Shorr et al, 2004). Chronic IEP is also rare, usually occurring in atopic women, and is characterised by peripheral pulmonary infiltrates (Marchand et al, 1998). Approximately 50% of patients have a history of asthma and the clinical phenotype includes chronic cough, malaise and weight loss. Chronic IEP responds well to systemic corticosteroid therapy but relapses are inevitable (Marchand et al, 1998).

Other miscellaneous causes of secondary eosinophilia

In October 1989, a group of physicians from New Mexico reported a group of patients, taking nutritional supplements containing l-tryptophan, with unusual visceral manifestations, accompanied by peripheral eosinophilia (Das et al, 2004). The patients manifested weakness, myalgia, arthralgia, rash, oral ulcers, alopecia, sclerodermiform skin changes and increased serum levels of muscle enzymes. This ‘eosinophilia-myalgia syndrome’ was subsequently reported in several hundred cases and associated with an increased risk of death (Sullivan et al, 1996). Similarly, in 1981, an outbreak of a deadly disease (toxic oil syndrome) appeared in Spain and was clinically characterised by severe myalgia, marked peripheral eosinophilia and pulmonary infiltrates (Diggle, 2001; Posada de la Paz et al, 2001; Sanchez-Porro Valades et al, 2003). Of the approximately 20 000 people affected, over 2500 deaths had occurred by December 1995 and epidemiological observations suggested a link with ingestion of adulterated rapeseed oil (Sanchez-Porro Valades et al, 2003).

Eosinophilic fasciitis (Schulman syndrome) is a scleroderma-like illness that was first described in 1974 and differentiated from scleroderma by the absence of Raynaud phenomenon and visceral involvement (Shulman, 1975; Mori et al, 2002). The disease is common in young adult males and the involved skin appears shiny and erythematous and subsequently becomes taut and woody with subsequent joint contractures. The pathogenesis of the disease is unknown and corticosteroids are used for treatment. Kimura disease is another rare chronic inflammatory disease that has a predilection for young males (Chen et al, 2004a). Clinical features include subcutaneous masses in the head and neck region associated with regional lymphadenopathy. Histological features include follicular hyperplasia, eosinophilic infiltrates and proliferation of postcapillary venules (Chen et al, 2004a). Wells syndrome is characterised by oedematous erythematous plaques that are pruritic and resolve promptly with systemic corticosteroid therapy (Falagas & Vergidis, 2005). In contrast, Omenn syndrome is a form of severe combined immunodeficiency associated with high mortality. The disease affects infants and is characterised by erythematous rash, hepatosplenomegaly, lymphadenopathy, recurrent infections and alopecia (Aleman et al, 2001). Outcome is fatal unless haematopoietic stem cell transplantation is performed.

Connective tissue/autoimmune diseases, especially systemic lupus erythematosus (Thomeer et al, 1999), polyarteritis nodosa (Kirkland et al, 1997) and scleroderma (Fleischmajer et al, 1978), as well as sarcoidosis (Renston et al, 2000), can be associated with mild eosinophilia. Other eosinophilia-associated chronic inflammatory conditions include inflammatory bowel disease (Benfield & Asquith, 1986) and chronic pancreatitis (Tokoo et al, 1992). The mechanism in the former instance might involve eotaxin-mediated chemotaxis (Garcia-Zepeda et al, 1996). The gastrointestinal system is a target organ for many eosinophilic disorders whereas organ-specific eosinophilic gastroenteritis might occur without associated blood eosinophilia (Khan, 2005). Finally, paraneoplastic eosinophilia is a well-known phenomenon in the setting of both metastatic cancer and lymphomas (Sataline & Mobley, 1967; Miller et al, 1977; Lowe & Fletcher, 1984; Balducci et al, 1989; Di Biagio et al, 1996; Scales & McMichael, 2001; Anagnostopoulos et al, 2005). On the other hand, the association of eosinophilia with either growth hormone or adrenal insufficiency is much less recognised (Spry, 1976; Kawada et al, 2001).

Clonal eosinophilia

The diagnosis of clonal eosinophilia requires the demonstration of either a cytogenetic/molecular marker of clonality or bone marrow histological features that are consistent with an otherwise classified myeloid malignancy. Examples of myeloid disorders that might be accompanied by clonal eosinophilia include both acute myeloid (AML) (Sanada et al, 1989) and lymphoblastic (ALL) (Blatt et al, 1974) leukaemias, chronic myeloid leukaemia (CML) (Keung et al, 2002), myelodysplastic syndrome (MDS) (Kuroda et al, 2000) and MPD (Bain, 2003).

Peripheral blood and bone marrow histological clues for clonal eosinophilia include macrocytosis, monocytosis, left shift granulocytosis, presence of circulating blasts, thrombocytosis, multilineage myeloproliferation, dyshaematopoiesis and reticulin fibrosis. However, bone marrow histological features can be subtle and the morphological distinction between clonal eosinophilia and idiopathic eosinophilia (including HES) is not always precise. In addition, intense bone marrow eosinophilia might make it difficult to identify neoplastic population of monocytes and mast cells. Therefore immunohistochemical stains for tryptase and mast cell immunophenotyping should accompany bone marrow examination in patients with an eosinophilic disorder before assigning a diagnosis of idiopathic eosinophilia. On the other hand, the detection of a clonal cytogenetic abnormality confirms the diagnosis of clonal eosinophilia regardless of how the bone marrow histology is interpreted. Furthermore, current evaluation of suspected HES mandates molecular investigation with either reverse transcription polymerase chain reaction (RT-PCR) or fluorescent in situ hybridisation (FISH) to exclude the possibility of the Gleevec®-sensitive, karyotypically occult Fip 1-like-1 (FIP1L1)/PDGFRA-positive eosinophilic disorder (Cools et al, 2003).

Molecular pathogenesis of clonal eosinophilia

Cytogenetic abnormalities in eosinophilic disorders are mostly non-specific but in certain instances have helped identify disease-causing mutations (Table V). Recently, recurrent molecular abnormalities have been identified in eosinophilia-associated MPD that have advanced our understanding of the molecular pathophysiology of these disorders and which increasingly support the development and use of a molecular classification for these heterogenous disorders. Excluding the molecularly well-characterised subtypes of AML, such as French–American–British (FAB) subtypes M4eo and M2 that exhibit eosinophilia, the chronic MPDs with associated eosinophilia are largely linked to constitutively active cellular tyrosine kinases, which drive the clonal cell proliferation. The clinical significance of the prospective identification of such mutant kinases is their susceptibility to molecularly targeted small-molecule inhibitors, such as Gleevec®, which frequently constitute a very effective treatment for these patients. Despite the identification of mutant tyrosine kinases, many questions regarding their role(s) in clonal eosinophilias remain unanswered to varying degrees (as illustrated below) – these pertain, for instance, to the varied genotype–phenotype association(s) and variable lineage distribution of specific molecular abnormalities, and to the role of additional molecular lesions and/or heritable genetic polymorphisms that influence the disease phenotype. Nevertheless, there is a remarkable association between activating mutations of certain receptor tyrosine kinases (e.g. PDGFR-α and -β) and eosinophilia-associated MPD. One possible explanation for this might relate to the common signalling pathways that such mutant kinases share with IL-5 and other eosinophilopoietic cytokines (Eriksson et al, 1992; Adachi & Alam, 1998; D'Andrea & Gonda, 2000; Paukku et al, 2000). However, a recent study has suggested a persistent role for IL-5 in the link between certain PDGFR mutations and eosinophilia, thus reflecting the complexity of the subject matter and the need for additional studies (Yamada et al, 2006).

Table V.   Cytogenetic anomalies reported in association with clonal eosinophilic disorders.
Chromosome affectedKaryotypeMolecular phenotypeClinicopathological presentationReferences
  1. HES, hypereosinophilic syndrome; CMPD, chronic myeloproliferative disorder; CEL, chronic eosinophilic leukaemia; MDS, myelodysplastic syndrome; EMS, eosinophilic mastocytosis; MS, mastocytosis; CML, chronic myeloid leukaemia; AML Eo, acute eosinophilic leukaemia; CMML, chronic myelomonocytic leukaemia; JMML, juvenile myelomonocytic leukaemia; NA, data not available; AT, acute transformation.

1t (1; 4) (q44:q12)
+1,dic (1; 7)(p10:q10)
1,der (1; 7)(q10; p10)
t (1; 5) (q21; q33)
t (1; 5) (q21; q31)
t (1; 5) (q21; q33)
t (1; 3; 5) (p36; p21; q33)
t (1; 5) (q23; p14)
Trisomy 1
PDGFRA-FIPILI rearrangement

Abnormality of PDGFRB

C-kit mutations
Atypical CML
Cools et al (2003)
Forrest et al (1998)
Park et al (2004)
Baxter et al (2003)
Baxter et al (2003)
Baxter et al (2003)
Baxter et al (2003)
Zermati et al (2003)
Harrington et al (1988)
2t (2; 4) (p24; q12)
t (2; 12; 5) (q37; q22; q33)
PDGFRB mutation
Musto et al (2004)
Musto et al (2004)
3t (3; 4) (p13; q12)
t (3; 5) (p21; q31)
t (3; 5) (p13; q13)
Atypical CML
Myint et al (1995)
Baxter et al (2003)
Shanske et al (1996)
4t (4; 7) (q11; q32)
t (4; 7) (q11; p13)
t (4; 16) (q11/12; p13)
del 4q12
t(4; 22) (q12; q11)

Atypical CML
Atypical CML
Atypical CML
Duell et al (1997)
Schoffski et al (2000)
Hild & Fonatsch, 1990)
Gotlib, 2005; Roche-Lestienne et al (2005)
Baxter et al (2002)
5t (5; 9) (q11; q34)
t (5; 11) (p15; q13)
t(5; 14) (q33; q24)
t (5; 17) (q33; p11)
t (5; 15) (q33; q22)
t (5; 12) (q33; p13)
t (5; 14) (q33; q32)

t(5; 7) (q33; q11·2)
t (5; 10) (q33; q11·2)
t (5; 17) (q33; p13)
t (5; 14) (q33; q32)
t(5; 16) (q33; q22)

Atypical CML
Atypical CML
AML after clonal evolution
Atypical CML
Bakhshi et al (2003)
Yoo et al (1984)
Vizmanos et al (2004a)
Morerio et al (2004)
Grand et al (2004a)
Golub et al (1994)
Abe et al (1997)
Ross et al (1998)
Kulkarni et al (2000)
Magnusson et al (2001)
Levine et al (2005)
Bhambhani et al (1986)
6t (6; 11) (q27; q23)
del (6) (q24)
t (6; 8) (p12; q12)
Suzuki et al (2001)
Gotlib (2005)
Popovici et al (1999)
7t (7; 12) (p11; q11)
Monosomy 7
da Silva et al (1988)
Humphrey et al (1981)
Song & Park (1987)
8t (8; 9) (p22; p23)
t (8; 9) (p21; p24)
Trisomy 8
+ 8 p23
+I (8p)
+8; +9
+8; + 21
t (8; 13) (p12; q12)
t (8; 9) (p12; q33)
t (8; 22) (p11; q11)
t (8; 17) (p11; q25)
t (8; 19) (p12; q13·3)

Vandenberghe et al (2004)
Weinfeld et al (1977)
Kook et al (2002)
Egesten et al (1997)
Mori et al (1986)
Kueck et al (1991)
Popovici et al (1998)
Guasch et al (2000)
Demiroglu et al (2001)
Sohal et al (2001)
Guasch et al (2003)
9Ins (9; 4) (934; q12q31) HESSchoch et al (2002)
10Trisomy 10
t(10; 11) (p14; q21)
Gotlib (2005)
Broustet et al (1986)
12Ins (12; 8) (p11; p11p22)FGFR1OP2-FGFR1EMSGrand et al (2004b)
15t (15; 21) (q13; q22)
+ 15
+ 15;− Y
Brito-Babapulle (1997)
Oliver et al (1998)
Weide et al (1997)
16t(16; 21) (p11; q22)
Inv (16) (p13q22)
t(16,16) (p13; q22)
Mecucci et al (1985)
Le Beau et al (1983)
Le Beau et al (1983)
17Isochromosome 17
Add (17) (q25)
Mitelman et al (1975)
Rotoli et al (2004)
20Del 20 (q11; q12) HESBrigaudeau et al (1996)
21Trisomy 21 HESKusanagi et al (1998)
Short Y
Y, del (15q22)
c-N-ras activationHES
Needleman et al (1990)
Flannery et al (1972)
Goffman et al (1983)
Complex cytogeneticsXYY, t (3; 5), +8,+mar
−21,+14,+11;del (5)q31
CEL with AT
Bitran et al (1977)
Ellman et al (1979)
Wolz et al (1993)
Bigoni et al (2000)
Cools et al (2004)

PDGFRA-rearranged eosinophilic disorders

Cytogenetically apparent: BCR-PDGFRA– t(4; 22)(q12; q11).  The first report of a chromosomal rearrangement t(4; 22)(q12; q11) targeting platelet-derived growth factor receptor (PDGFR)-α (PDGFRA) pre-dated Gleevec® use (Baxter et al, 2002). Molecular studies on two patients presenting with atypical CML with associated eosinophilia (one patient had extra-medullary T-lymphoblastic lymphoma concurrently at diagnosis) revealed an in-frame breakpoint cluster region (BCR) (breakpoints in intron 7 and exon 12)-to-PDGFRA (breakpoints in exon 12) fusion mRNA. Subsequently, Trempat et al (2003) described a case with atypical CML evolving into pre-B ALL, with the BCR (exon 1)-to-PDGFRA (exon 13) fusion, who achieved a complete haematological remission with Gleevec® treatment. Finally, Safley et al (2004) reported a case with atypical CML and eosinophilia, with the BCR (exon 17)-to-PDGFRA (exon 12) fusion, who also responded to Gleevec®. All cases exhibited the t(4; 22)(q12; q11) cytogenetic abnormality, and the PDGFRA breakpoints were noted to be tightly clustered in the juxtamembrane (JM) region, pointing to a key regulatory (auto-inhibitory) role for this domain. Similar targeting of the JM domain is also seen in the other PDGFRA-mediated diseases, namely HES with FIP1L1-PDGFRA (Cools et al, 2003) (see below) and gastrointestinal stromal tumours (Heinrich et al, 2003). Another study (Ketterling et al, 2004), screened 29 047 archived abnormal bone marrow karyotypes and found 11 cases with a breakpoint involving PDGFRA (Ketterling et al, 2004). Three cases implicate novel translocation partners (1q44, 3q25 and 17q23) that have not been cloned as of yet.

Cytogenetically occult: FIP1L1-PDGFRA.  Following the remarkable success of Gleevec® at a relatively low-dose (∼100 mg/day) in the treatment of HES [summarised in (Pardanani & Tefferi, 2004; Gotlib, 2005)], a concerted multi-institutional effort identified the FIP1L1-PDGFRA tyrosine kinase as the molecular target for imatinib in a subset of HES patients (Cools et al, 2003). Cloning of the FIP1L1-PDGFRA fusion gene identified a novel molecular mechanism for generating this constitutively active fusion tyrosine kinase, wherein a c. 800 kb interstitial deletion within 4q12 fuses the 5′ portion of FIP1L1 to the 3′ portion of PDGFRA (Cools et al, 2003). Molecular studies show that the breakpoint in FIP1L1 is relatively promiscuous, while the PDGFRA breakpoint is restricted to exon 12 that encodes part of the protein–protein interaction module with two fully conserved tryptophans (WW domain)-containing JM region (Cools et al, 2003; Roche-Lestienne et al, 2005). Given the known auto-inhibitory role of the JM region for other receptor tyrosine kinases including EPHB2 (Wybenga-Groot et al, 2001), FLT3 (Gilliland & Griffin, 2002) and KIT (Chan et al, 2003), disruption of this domain is probably the primary mechanism for FIP1L1-PDGFRA activation.

Since its original description (Cools et al, 2003), other studies have begun to clarify the prevalence and clinicopathological associations of the FIP1L1-PDGFRA mutation (summarised in Table VI). Our experience suggests that in unselected cases of eosinophilia, prevalence of the mutation is quite low (c. 4%) (Pardanani et al, 2005), but this remains to be confirmed in a multi-institutional setting. Prevalence of the mutation is higher in cohorts that satisfy World Health Organization (WHO) criteria (Bain et al, 2001) for idiopathic HES (12–88%), and particularly for the subgroup with myeloproliferative features, including marrow fibrosis, an elevated serum tryptase level and increased numbers of neoplastic mast cells in the bone marrow. These patients have been labelled as myeloproliferative variant of HES (HES-MPD or MP-HES) (Klion et al, 2003, 2004b) or as eosinophilia associated-systemic mast cell disease (SM-CEL) (Pardanani et al, 2003a, 2004), a distinction that is therapeutically irrelevant, given that presence of the fusion predicts Gleevec® responsiveness. Fourteen of the 15 FIP1L1-PDGFRA+ patients (93%), of over 220 patients tested at our institution, have been found to have SM-CEL/HES-MPD (one patient had CEL by WHO criteria). This data suggests that the phenotypic spectrum associated with the FIP1L1-PDGFRA mutation, at least as detected by our FISH methodology, is quite narrow and may be largely restricted to SM-CEL/HES-MPD. The latter point will need to be confirmed through multi-institutional effort that employs uniform diagnostic criteria, including measurement of serum tryptase levels and expert examination of bone marrow sections using stains that highlight the relatively subtle mast cell infiltrates in such cases (Pardanani et al, 2004; Pardanani, 2005).

Table VI.   Summary of FIP1L1-PDGFRA studies, including imatinib-response data.
ReferenceTotal number of cases (male)Eligibility (number)Method (number tested for FIP1L1- PDGFRA)FIP1L1- PDGFRA positive casesTotal imatinib treatedImatinib dose (positive cases)Imatinib response (positive cases)Imatinib dose (negative cases)Imatinib response (negative cases) (%)Other abnormalities
  1. *One additional patient (FIP1L1-PDGFRA mutation status unknown) also had complete remission to imatinib therapy.

  2. n, number; HES, idiopathic hypereosinophilic syndrome; WHO, World Health Organization; CEL, chronic eosinophilic leukaemia; Ph, Philadelphia chromosome; AML, acute myeloid leukaemia; SMCD, systemic mast cell disease; ALL, acute lymphoblastic leukaemia; SMCD-HES, eosinophilia associated-SMCD; nos, not otherwise specified; RT-PCR, reverse transcription-polymerase chain reaction; FISH, fluorescence in situ hybridisation; NG, not given; NA, not applicable; CHR, complete haematological response; PR, partial response; PDGFRA, platelet-derived growth factor receptor (PDGFR)-α; PDFRB, PDGFR-β; FGFR1, fibroblast-derived growth factor receptor 1.

Pardanani et al (2005)830 (NG)Unselected patients with eosinophilia and/or SMCDFISH (all)32 (4%)11/14 (79%)100–400 mg/d11/11 (100%) CHRNANANA
Roche-Lestienne et al (2005)35 (23)HES/WHO + normal cytogeneticsRT-PCR (n = 35)
FISH (n = 29)
6/35 (17%)9/35 (26%)100–200 mg/d5/5 (100%) CHR100–200 mg/d1/4 (25%) CHRT-cell clonality 11/35 (31%)
La Starza et al (2005)26 (17)HES/WHO (n = 20); non-secondary eosinophilia > 1·5 × 109/l (n = 6)FISH (n = 26)
RT-PCR (n = 4)
10/26 (38%)15/26 (58%)100–400 mg/d7/7 (100%) CHR200–600 mg/d2/5 (20%) CHRPDFRB, FGFR1, ETV6, ABL1 – not rearranged
Vandenberghe et al (2004)17 (13)HES or CEL/WHO (clonal T-cell cases excluded)RT-PCR (n = 17)
FISH (n = 10)
8/17 (47%)5/17 (29%)100 mg/d4/4 (100%) CHR100 mg/d0/1NA
Pardanani et al (2004)89 (NG)Eosinophilia > 1·5 × 109/l (Ph+ and t(5; 12)(q33; p13) excluded)FISH (n = 89)11/89 (12%)26/89 (29%)100–400 mg/d8/8 (100%) CHR100–400 mg/d4/18 (22%) PRNA
Cools et al (2003)17 (13)HES/WHO (n = 16); AML (n = 1)RT-PCR (n = 17)
FISH (n = 1)
9/16 (56%)11/17 (65%)100–400 mg/d5/5 (100%) CHR100–400 mg/d4/5 (80%) CHRPDGFRA, PDGFRB, KIT – no mutations
Klion et al (2003)32 (16)HES/WHO (n = 15); helminth (n = 6); SMCD (n = 3); other (n = 8)RT-PCR (n = 11)5/11 (45%)6/32 (19%)400 mg/d5/5* (100%) CHRNANAAll three SMCD cases were KIT D816V+
Klion et al (2004b)8 (8)MHES ‘myeloproliferative variant’ of HES (n = 7); ALL (n = 1)RT-PCR (n = 8)7/8 (88%)7/8 (88%)400 mg/d7/7 (100%) CHRNANANA
Pardanani et al (2003a)5 (NG)SMCD-HESFISH (n = 5)
RT-PCR (n = 1)
3/5 (60%)5/5 (100%)100–400 mg/d3/3 (100%) CHR100–400 mg/d0/22/5 cases were KIT D816V+; none had Ph, t(5; 12)(q33; p13) or KIT or PDGFRB mutations
Martinelli et al (2004)55 (NG)HES/WHORT-PCR (n = 55)13/55 (24%)31/55 (56%)100–400 mg/d13/13 (100%) CHR100–400 mg/d1/18 (6%) PRPDGFRB-TEL, FGFR1-BCR and BCR-ABL not detected
Schoch et al (2004)40 (27)Eosinophilia nosFISH (n = 40)
Chromosome banding analysis (n = 37)
RT-PCR (n = 4)
4/40 (10%)NANANANANAClonal cytogenetic abnormalities noted in six patients (including one with FIP1L1-PDGFRA)

All individuals carrying the FIP1L1-PDGFRA mutation achieve a complete haematological remission with 100–400 mg/d of Gleevec®, which is considered first-line treatment for this cohort (summarised in Table VI). The drug is taken with a full glass of water, once a day and always with meals to avoid upper gastrointestinal irritation. Drug side effects include periorbital and peripheral oedema, diarrhoea, nausea, muscle cramps, fatigue, bone pain and rash. Peripheral blood screening for FIP1L1-PDGFRA, using either FISH or RT-PCR, can be performed to monitor molecular response to treatment at 3–6 month intervals in the first year and at 6–12 months interval afterwards. The cumulative data indicates that, after initial induction therapy, most FIP1L1-PDGFRA+ patients [barring a small minority (Klion et al, 2004b; Vandenberghe et al, 2004)], achieve molecular remission within weeks to months of starting Gleevec® therapy, regardless of whether nested RT-PCR (Klion et al, 2004b; Martinelli et al, 2004; Vandenberghe et al, 2004; La Starza et al, 2005; Roche-Lestienne et al, 2005) or interphase FISH (Pardanani et al, 2003a, 2004; La Starza et al, 2005) is used as the monitoring tool. Given the short follow up of published reports and the rarity of such cases, it is currently unknown whether patients who fail to achieve a molecular remission have a different natural history as compared to those achieving such a remission. For the former group, is it possible that, similar to BCR-ABL in CML (Roche-Lestienne et al, 2002, 2003; Deininger et al, 2005), resistance-inducing mutations in the PDGFRA kinase domain predate initiation of Gleevec® therapy and may be present at diagnosis? While a FIP1L1-PDGFRA mutation (T674I) that is homologous to the resistance-inducing, ‘gatekeeper’ T315I mutation in BCR-ABL has been described for two cases of FIP1L1-PDGFRA+ CEL in blast crisis (both in the setting of an aberrant karyotype) (Cools et al, 2003; von Bubnoff et al, 2005), the issue of whether this mutation arises without selective pressure from Gleevec® has not been studied thus far. Whether the FIP1L1-PDGFRA T674I mutation remains sensitive to higher doses of Gleevec®, as do some BCR-ABL mutations (Corbin et al, 2003), is currently unknown.

PDGFRB-rearranged eosinophilic disorders

Golub et al (1994) first reported fusion of the tyrosine kinase encoding region of PDGFRB to the ets-like gene, ETV6 (previously known as tel) in a patient with chronic monomyelocytic leukaemia (CMML) and the t(5; 12)(q33; p13) cytogenetic abnormality. Since then, other fusion transcripts have been cloned, wherein PDGFRB is fused to the N-terminal segment of a partner protein that encodes for one or more oligomerisation domains (Table VII) [summarised in (Pardanani & Tefferi, 2004; Gotlib, 2005)]. These patients carry 5q31–33 chromosomal rearrangements and generally present as atypical CML or a hybrid MPD/MDS syndrome (CMML, juvenile monomyelocytic leukaemia, etc.), frequently with associated eosinophilia. Translocation t(5; 12), however, is a relatively rare abnormality. A Mayo Clinic review of 56 709 cases identified only 25 such cases (0·04%). Of the 11 patients for whom clinical data was available, only three had eosinophilia. Further, in a cohort of 213 CMML patients, of which 205 had karyotype analysis, none were found to carry t(5; 12), even though 34% had other cytogenetic abnormalities (Onida et al, 2002). Importantly, as shown elegantly in a study by Baxter and colleagues, for patients carrying translocations involving 5q31–33, the mere finding of a 5q33 breakpoint (where PDGFRB is assigned) in patients with a myeloid disorder does not necessarily indicate that PDGFRB is involved (Baxter et al, 2003). Conversely, involvement of 5q31 does not exclude PDGFRB involvement given that translocations may be complex at the molecular level. Hence, molecular studies are essential in patients with 5q31–33 translocations to confirm or exclude PDGFRB rearrangement, given the predictive value for Gleevec® responsiveness of such a finding. Other novel translocations involving PDGFRB for which the partner genes remain to be identified have been reported (Table VII) (Baxter et al, 2003; Ketterling et al, 2004). As summarised in Table VII, Gleevec® therapy has in most cases resulted in complete haematological and/or cytogenetic remissions in patients with PDGFRB gene fusions.

Table VII.   Summary of reports describing PDGFRB-fusion genes, including imatinib-response data.
ReferenceTotal number of cases (male)Clinical presentationCytogenetics data providedPDFGRB fusion partnerReciprocal transcript detected?Imatinib doseImatinib responseOther pertinent references
  1. CML, chronic myeloid leukaemia; aCML, atypical (Philadelphia chromosome-negative) CML; MDS, myelodysplastic syndrome; MPD, myeloproliferative disorder; CMML, chronic myelomonocytic leukaemia; JMML, juvenile myelomonocytic leukaemia; AML, acute myeloid leukaemia; NG, not given; NA, not applicable; CH, complete haematological response; CC, complete cytogenetic response; PH, partial haematological response; PDGFRB, platelet-derived growth factor receptor-β.

Grand et al (2004a)1 (1)• aCML, eosinophilia• t(5; 15)(q33; q22)p53BP1No300–400 mg/dPHR 
Vizmanos et al (2004a)1 (1)• aCML, eosinophilia• t(5; 14)(q33; q24)NINYes200–400 mg/dCHR/CCR 
Morerio et al (2004)1 (1)• JMML, eosinophilia• 46,XY,t(5; 17)(q33; p11·2)HCMOGT-1NoNANA 
Wilkinson et al (2003)1 (0)• MDS/MPD, eosinophilia• t(1; 5)(q23; q33)PDE4DIP (myomegalin)NoNGCHR 
Golub et al (1994)1 (?)• CMML• t(5; 12)(q33; p13)ETV6 (TEL)NoNANAWlodarska et al (1995)
Apperley et al (2002)4 (4)• chronic MPD, eosinophilia• 46,XY,t(5; 12)(q33; p13)ETV6 (TEL) (3 of 4 patients)NA400–800 mg/dCHR/CCRPitini et al (2003b)
Kulkarni et al (2000)1 (1)• aCML, eosinophilia• 46,XY,t(5; 10)(q33; q21·2)H4/D10S170NoNANASchwaller et al (2001)
Garcia et al (2003)1 (1)• aCML, eosinophilia• 46,XY,t(5; 10)(q33; q22)H4/D10S170NG400 mg/dCHR/CCR 
Ross et al (1998)1 (1)• CMML, eosinophilia• t(5; 7)(q33; q11·2)HIP1NoNANA 
Magnusson et al (2001)1 (1)• CMML• 46,XY,t(5; 17)(q33; p13)Rabaptin-5NoNANA 
Abe et al (1997)1 (0)• AML, eosinophilia• 46,XX,t(5; 14)(q33; q32), t(7; 11)(p15; p15)CEV14NoNANA 
Baxter et al (2003)1 (1)• aCML, eosinophilia• t(3; 5)(p21; q31)• Not identifiedNANANA 
Kim et al (2005)1 (?)• CMML, eosinophilia• t(5; 12)(q31; p13), t(1; 7)(q10; p10)• Not identifiedNANANA 
Ketterling et al (2004)3 (NG)• NG• t(1; 5)(q21; q33)
• t(5; 14)(q33; q32)
• t(5; 16)(q33; p13·1)
• Not identifiedNANANA 

FGFR1-rearranged eosinophilic disorders

The rearrangement, and consequent activation, of fibroblast growth factor receptor 1 (FGFR1) is associated with a syndrome known as the 8p11 myeloproliferative syndrome (EMS) or stem cell leukaemia lymphoma syndrome (SCLL). EMS is an aggressive myeloproliferative disorder frequently associated with eosinophilia and T-cell lymphoblastic lymphoma (Macdonald et al, 2002). Both myeloid and lymphoid lineage cells exhibit the 8p11 translocation, thus demonstrating the stem cell origin of this disease. Clinically, a biphasic course is frequently observed – a relatively short chronic phase, followed by transformation into acute leukaemia with a poor overall prognosis. While intensive chemotherapy with allogeneic stem cell rescue is considered the only potential curative therapy for EMS, the use of FGFR1-targeting small molecule kinase inhibitors such as protein kinase C (PKC) 412 (an N-benzoyl derivative of the naturally occurring alkaloid staurosporine) for treating EMS patients is currently being investigated (Chen et al, 2004b). Similar to PDGFRB, the FGFR1 fusion proteins are constitutively active tyrosine kinases, wherein the N-terminal partner protein contributes self-association domain(s). Following ZNF198-FGFR1, numerous other chimaeric genes resulting from FGFR1 rearrangement, all with an exon 9 breakpoint, have been identified to date (summarised in Table VIII). Of note, Roumiantsev et al (2004) have modelled EMS and atypical CML (resembling human disease) in mice using ZNF198-FGFR1 and BCR-FGFR1 fusion constructs, respectively, in the murine bone marrow transduction/transplantation system. In the EMS model, the FGFR1 Y766F mutation was found to attenuate both myeloid and lymphoid diseases, thus implicating phospholipase C that disrupts Grb2 binding was found to cause EMS-like disease. These data implicate different signalling pathways originating from both the FGFR1 kinase as well as the fusion partner in the pathogenesis of atypical CML and EMS. Further, these mouse models potentially serve as a platform for testing drugs that target dysregulated FGFR1, such as PKC412.

Table VIII.   Summary of reports describing FGFR1-fusion genes.
ReferenceTotal number of cases (male)Clinical presentationCytogenetics data providedFGFR1 fusion partnerReciprocal transcript detected?Other pertinent references
  1. FGFR1, fibroblast-derived growth factor receptor 1; aCML, atypical (Philadelphia chromosome-negative) chronic myeloid leukaemia; MPD, myeloproliferative disease; AML, acute myeloid leukaemia; PV, polycythaemia vera; B-ALL, B-cell acute lymphoblastic leukaemia; SMCD, systemic mast cell disease; NG, not given; NA, not applicable.

Walz et al (2005)1 (0)aCML, eosinophilia, basophilia, monocytosis47,XX,t(8:17)(p11; q23),+20MYO18ANo 
Belloni et al (2005)1 (0)AML-M4, eosinophilia, monocytosist(7; 8)(q34; p11)TIF1Yes 
Grand et al (2004b)1 (1)T-lymphoblastic lymphoma, eosinophilia→AMLins(12; 8)(p11; p11p12)FGFR1OP2No 
Guasch et al (2003)1 (1)AML-M0, eosinophilia45,X,t(8; 19)(p12; q13·3),−YHERV-KNo 
Guasch et al (2000)1 (1)MPD, eosinophilia→AML46,XY,t(8; 9)(p12; q33){8}/48, idem,+der(9)t(8; 9);+21{12}CEP110Yes 
Demiroglu et al (2001), Fioretos et al (2001)• 1 (1)
• 2 (0)
• MPD, eosinophilia, basophilia→AML
• aCML, eosinophilia, basophilia
• 46,XY,t(8; 22)(p11; q11), ?dup(9)(q34; q34)
• t(8; 22)(p11; q11)
BCRYesPini et al (2002), Murati et al (2005)
Popovici et al (1999)2 (2)• MPD, eosinophilia→ PV→AML,
46,XY,t(6; 8)(q27; p11)FOPYesVizmanos et al (2004b)
Popovici et al (1998)2 (NG)• T-lmyphoblastic lymphoma/MPD eosinophilia→AML
• 48,XX,t(8; 13)(p12; q12), +der(13)t(8; 13)(p12; q12), +19{4}/51,idem,+6,+der(8),t(8; 13) (p12; q12),+der(13),t(8; 13)(p12; q12){2}
• t(8; 13)(p12; q12)
FIM (synonymous with ZNF198)Yes 
Smedley et al (1998)2 (NG)T-lmyphoblastic lymphoma/MPD, eosinophilia t(8; 13)(p11; q11-12)RAMP (synonymous with ZNF198)No 
Xiao et al (1998)4 (NG)? t(8; 13)(p11; q11-12)ZNF198? 
Reiter et al (1998)5 (NG)MPD/lymphoma, eosinophilia (individual patient details not provided) t(8; 13)(p11; q12)ZNF198NoMatsumoto et al (1999)
Sohal et al (2001)• 1 (1)
• 1 (1)
• 1 (0)
• 1 (1)
• T-lymphoma/MPD, eosinophilia
• T-lymphoma/MPD, eosinophilia
• 46,XY,t(8; 17)(p11; q25)
• 47,XY,t(8; 11)(p11; p15),+8,−17,
• 46,XX,t(8; 12)(p11; q15)
• 46,XY,ins(12; 8)(p11; p11p21)
FGFR1 predicted to be disrupted in all four cases
FGFR1 partner genes not identified

Idiopathic eosinophilia

Once secondary eosinophilia is considered unlikely (Tables I–IV), a working diagnosis of primary eosinophilia is made and a bone marrow biopsy with appropriate cytogenetic, molecular, histochemical and immunophenotypic studies is recommended in order to characterise the process further as either clonal or idiopathic eosinophilia (see above discussion on clonal eosinophilia). HES is a subset of idiopathic eosinophilia that fulfils the traditional criteria of a persistent (>6 months) increase in AEC (>1·5 × 109/l) associated with target organ damage (Chusid et al, 1975). However, evidence from both X-linked clonality studies (Chang et al, 1999; Malcovati et al, 2004) and long-term follow up of affected patients suggest that at least some patients with HES harbour an underlying clonal myeloid malignancy that could progress into frank acute leukaemia or MPD (Owen & Scott, 1979; Yoo et al, 1984; Brown & Stein, 1989; Needleman et al, 1990). On the other hand, the demonstration of a clonal (Simon et al, 1999) or phenotypically abnormal (Simon et al, 1999) T-cell population in other patients with HES suggests the alternative possibility of a true secondary process with some of these patients progressing into clinically overt lymphoma (Butterfield, 2001).

Clinical manifestations

As is the case with clonal eosinophilia, over 90% of patients with HES are males and the disease is rare in children (Chusid et al, 1975; Yildiran & Ikinciogullari, 2005). Clinical manifestations are markedly heterogeneous and the disease can either be completely asymptomatic or involve multiple organs including the skin (pruritus, urticaria, angioedema, erythematous papules or nodules, mucosal ulcers), the heart (fibroblastic endocarditis, valvular disease, mural thrombi, cardiomyopathy, elevated troponin levels), the nervous system [sensorimotor polyneuropathies, mononeuritis multiplex, isolated central nervous system (CNS) vasculitis, optic neuritis, acute transverse myelitis], the lung (pulmonary infiltrates, lung nodules, pleural effusion), the gastrointestinal system (hepatosplenomegaly, gastroenteritis, sclerosing cholangitis), the haematopoietic system (cytopenias, bone marrow fibrosis) and the kidney (thrombotic microangiopathy) (Fauci et al, 1982; Leiferman et al, 1982; Harley et al, 1983; Moore et al, 1985; Lefebvre et al, 1989; Weller & Bubley, 1994; Ommen et al, 2000; Liapis et al, 2005). In essence, therefore, any organ is vulnerable to eosinophilia-associated tissue damage although the major tissue targets are the heart, the nervous system, the skin and the upper and lower respiratory tract including the lung parenchyma. In addition, thromboembolic disease involving the cardiac chambers (Kocaturk & Yilmaz, 2005) as well as both venous (Liao et al, 2005) and arterial (Ponsky et al, 2005) vessels are not infrequent.

Regarding clinical course, it is important to remember that HES is a potentially fatal disease with a less than 50% reported 10-year survival (Lefebvre et al, 1989), especially in corticosteroid-resistant cases with cardiac involvement. However, future survival figures are expected to be better because FIP1L1/PDGFRA-positive cases are no longer considered as HES and their inadvertent inclusion in older studies (pre-Gleevec® era) probably contributed to the reported poor survival (Lefebvre et al, 1989). Similarly, certain clinical presentations, including recurrent or persistent angioedema and increased serum IgE levels, have been associated with the female gender and a more indolent clinical course free of cardiac involvement (Gleich et al, 1984; Chikama et al, 1998).

Laboratory evaluation

In addition to bone marrow biopsy and cytogenetic studies, evaluation of primary eosinophilia should include serum tryptase (an increased level suggests SM-CEL and warrants bone marrow histochemical studies for tryptase, mast cell immunophenotyping and molecular studies to detect either FIP1L1/PDGFRA or KITD816V) (Pardanani et al, 2004), T-cell immunophenotyping as well as T-cell receptor antigen gene rearrangement analysis (a positive test suggests an underlying clonal or phenotypically abnormal T-cell disorder and warrants measurement of IL-5 as well as consideration of T-cell-directed therapy) (Butterfield, 2001), serum IL-5 (an elevated level requires careful evaluation of the bone marrow as well as the T-cell gene rearrangement studies for the presence of a clonal T-cell disease and treatment with interferon-α might be considered because of the drug's effect on down-regulating IL-5 production by Th2 cells) (Schandene et al, 1996; Butterfield, 2001; Simon et al, 2001) and serum IgE level (patients with increased IgE level might respond better to corticosteroids and be at a lower risk of developing eosinophilia-associated heart disease) (Bush et al, 1978; Gleich et al, 1984).

Initial evaluation of the patient with primary eosinophilia should also include laboratory tests to look for eosinophilic-mediated tissue damage. In apparently asymptomatic patients, these include echocardiogram, chest X-ray, pulmonary function tests and measurement of serum troponin levels. Increased level of serum cardiac troponin has been shown to correlate with the presence of cardiomyopathy in HES and recent studies have suggested a predictive role for drug-induced cardiogenic shock during treatment with Gleevec® (Sato et al, 2000; Pitini et al, 2003a). In symptomatic patients, tissue biopsy might be required but not always essential to document causality.


There is currently no consensus regarding the management of asymptomatic patients with HES with no evidence of organ damage. One can argue instituting specific therapy, even in such cases, to prevent long-term ill effects from chronic organ exposure to excess eosinophils. However, there is no systematic study that supports such a concern and long term drug therapy has its own potential danger. Therefore, we currently prefer to closely monitor rather than to treat asymptomatic patients, regardless of the degree of eosinophilia. Accordingly, we recommend measurement of serum troponin level every 3–6 months and an echocardiogram every 6–12 months.

For the treatment of symptomatic patients with HES, the first-line drug of choice is prednisone (starting dose of 1 mg/kg/d) because of the rapidity and reliability of its effect. However, despite a near 70% overall response rate (Parrillo et al, 1978), relapses off therapy are usual and either a substitute drug or a steroid-sparing agent soon becomes necessary. In this regard, interferon-α (starting dose 3 million units three times a week) (Butterfield & Gleich, 1994; Ceretelli et al, 1998; Yoon et al, 2000; Baratta et al, 2002) and hydroxyurea (starting dose 500 mg twice a day) (Parrillo et al, 1978) have respectively served these roles by producing remissions in the majority of treated patients and are currently considered second-line drugs of choice. In true HES (i.e. FIP1L1/PDGFRA-negative), low-dose Gleevec® (100 mg/d) is unlikely to produce durable complete remissions (Pardanani et al, 2004). A higher dose of the drug (400 mg/d), however, might induce partial remissions (Pardanani et al, 2004) and in some instances a complete remission (Cools et al, 2003), thus making Gleevec® a reasonable third-line drug of choice. During Gleevec® therapy for patients with HES, it is important to recognise the possibility of drug-induced acute cardiac shock (Pardanani et al, 2003b; Pitini et al, 2003a) as well as treatment-associated oligospermia (Seshadri et al, 2004). The former is managed by the concomitant use of systemic corticosteroid therapy, which is recommended in the presence of either elevated serum troponin level or an abnormal echocardiogram (Pitini et al, 2003a).

In patients that are refractory to usual therapy in HES, treatment agents that have been used with some efficacy include chlorambucil (Weller & Bubley, 1994), etoposide (Smit et al, 1991), cyclosporine (Nadarajah et al, 1997), vincristine alone(Spry, 1982) or in combination with mercaptopurine (Marshall & White, 1989), cladribine (2-chlorodeoxyadenosine) alone (Ueno et al, 1997) or in combination with cytarabine (Ueno et al, 1997; Jabbour et al, 2005) and combination of cytarabine and 6-thioguanine (Eakin et al, 1982). Most recently, two monoclonal antibodies were evaluated; mepolizumab (SB 240563) targets IL-5 and alemtuzumab (Campath®) targets the CD52 antigen that is expressed by eosinophils but not neutrophils. Both were effective in controlling blood eosinophilia as well as disease symptoms. However, while durable remissions were seen with maintenance therapy with alemtuzumab (30 mg every 3 weeks) (Pitini et al, 2004; Sefcick et al, 2004), response to single-dose mepolizumab therapy (1 mg/kg) was relatively short lived and associated with rebound eosinophilia (Koury et al, 2003; Plotz et al, 2003; Kim et al, 2004; Klion et al, 2004c). Amongst the aforementioned drug options for salvage therapy, we prefer to use single agent rather than combination chemotherapy and avoid the use of alkylating agents. Otherwise, additional treatment experience is needed to enable choosing one of the remaining agents over the other. Finally in drug-refractory HES, myeloablative and non-myeloablative allogenic peripheral blood stem cell transplants have been used and were found to reverse organ dysfunction (Juvonen et al, 2002; Ueno et al, 2002; Cooper et al, 2005).


The serendipitous observation of Gleevec® activity in a subset of patients with systemic mastocytosis associated with eosinophilia (SM-CEL) (Pardanani et al, 2003c), whose clinical phenotype might be difficult to distinguish from that of HES (Gleich et al, 2002), has led to the discovery of FIP1L1/PDGFRA (Cools et al, 2003) as both a disease-causing mutation as well as marker of Gleevec® sensitivity. Accordingly, FIP1L1/PDGFRA has now joined a group of oncogenic kinases, both cytoplasmic and receptor tyrosine kinases, that are associated with MPD, including BCR/ABL, Janus kinase 2 (JAK2)V617F, KITD816V and others (Cross & Reiter, 2002; De Keersmaecker & Cools, 2005; Tefferi & Gilliland, 2005).

The myeloproliferation-inducing property of mutant tyrosine kinases is consistent with the role of their wild-type counterparts as relay points of signal transmission for haematopoietic growth factors. Furthermore, like AML, clonal myeloproliferation in MPD might represent a multistep process of multiple mutations that are individually responsible for cell proliferation/impaired apoptosis (e.g. mutations involving signal molecules), blockage of cell differentiation (e.g. mutations involving transcription factors) and overt leukaemic transformation (e.g. mutations involving tumour suppressor genes) (Frohling et al, 2005; Reilly, 2005). Such a contention is supported by the inverse correlation between Gleevec® treatment efficacy and disease duration/stage in CML (Goldman, 2004) as well as the lack of correlation between leukaemic transformation in MPD and JAK2V617F mutational status (Tefferi et al, 2005; Wolanskyj et al, 2005). Regardless, it is reasonable to expect retardation of disease progression in mutant kinase-driven MPD by the early institution of specific kinase inhibitor therapy.

Controversies aside (Cools et al, 2003), true HES remains molecularly undefined and not durably responsive to treatment with Gleevec® (Pardanani et al, 2004). Currently recognised Gleevec®-sensitive mutant kinase targets include BCR/ABL, FIP1L1/PDGFRA, fusion kinases involving PDGFRB and KIT mutations other than KITD816V (Pardanani & Tefferi, 2004). Accordingly, Gleevec® is also ineffective in FIP1L1/PDGFRA-negative SM-CEL (Pardanani et al, 2003a). Most recently, the aforementioned Gleevec®-resistant KITD816V has displayed in vitro treatment sensitivity to other oral kinase inhibitors including dasatinib (BMS-354825) (Schittenhelm et al, 2004; Shah et al, 2004) and PKC-412 (Gleixner et al, 2005). Dasatinib is an ATP-competitive, dual SRC/ABL inhibitor that is greater than 300-fold more potent than Gleevec® against BCR/ABL-transduced cells and has demonstrated preclinical activity against 18 of 19 Gleevec®-resistant BCR-ABL mutations. PKC-412 (N-benzoyl-staurosporine) is an indolocarbazole staurosporine analogue, which competes for binding to the ATP site on PKC family of serine–threonine kinases. Both dasatinib and PKC-412 are currently undergoing clinical trials in systemic mastocytosis (H. Kantarjian, personal communication). On the other hand, effective targeted therapy in ‘HES’ awaits additional insight in the molecular pathogenesis of the disease, unless serendipity strikes again.