Clinical and molecular features of FIP1L1-PDFGRA (+) chronic eosinophilic leukemias


Detection of the FIP1L1-PDGFRA fusion gene or the corresponding cryptic 4q12 deletion supports the diagnosis of chronic eosinophilic leukemia (CEL) in patients with chronic hypereosinophilia. We retrospectively characterized 17 patients fulfilling WHO criteria for idiopathic hypereosinophilic syndrome (IHES) or CEL, using nested RT-PCR and interphase fluorescence in situ hybridization (FISH). Eight had FIP1L1-PDGFRA (+) CEL, three had FIP1L1-PDGFRA (−) CEL and six had IHES. FIP1L1-PDGFRA (+) CEL responded poorly to steroids, hydroxyurea or interferon-α, and had a high probability of eosinophilic endomyocarditis (n=4) and disease-related death (n=4). In FIP1L1-PDGFRA (+) CEL, palpable splenomegaly was present in 5/8 cases, serum vitamin B12 was always markedly increased, and marrow biopsies revealed a distinctively myeloproliferative aspect. Imatinib induced rapid complete hematological responses in 4/4 treated FIP1L1-PDGFRA (+) cases, including one female, and complete molecular remission in 2/3 evaluable cases. In the female patient, 1 log reduction of FIP1L1-PDGFRA copy number was reached as by real-time quantitative PCR (RQ-PCR). Thus, correlating IHES/CEL genotype with phenotype, FIP1L1-PDGFRA (+) CEL emerges as a homogeneous clinicobiological entity, where imatinib can induce molecular remission. While RT-PCR and interphase FISH are equally valid diagnostic tools, the role of marrow biopsy in diagnosis and of RQ-PCR in disease and therapy monitoring needs further evaluation.


Idiopathic hypereosinophilic syndrome (IHES) and chronic eosinophilic leukemia (CEL) are rare chronic disorders with eosinophil overproduction in the bone marrow, which results in marked and sustained peripheral blood eosinophilia. Eventually, this leads to eosinophilic infiltration and functional damage of peripheral organs.1,2,3 The diagnostic criteria for IHES or CEL, as defined by the WHO, include: (1) eosinophilia exceeding 1.5 × 109/l, and persisting for more than 6 months; (2) exclusion of known causes of eosinophilia, including clonal or abnormal T-cell populations; and (3) signs and symptoms of organ involvement.4 The WHO recommendation to diagnose CEL in case of proven clonality, and IHES in the remaining cases,4 remains somewhat problematic in clinical practice, as cytogenetic aberrations or elevated peripheral or central blastosis are not readily identified in the majority of patients.1,3 Moreover, proposed surrogate markers of CEL, for example, elevated serum vitamin B12 levels, hepatomegaly or splenomegaly, do not allow a reliable distinction from IHES.5,6 Finally, organ damage, most frequently of the heart, can occur in CEL, in IHES and in eosinophilias caused by multiple other etiologies. Endocardial damage, eosinophil myocardial infiltration and degranulation represent the earliest lesions. They are followed by the formation of ventricular thrombi, typically sparing the aortic or pulmonic outflow tracts. Finally, progressive scarring with entrapment of chordae tendineae, mitral and/or tricuspid valve regurgitation, and endomyocardial fibrosis leads to a restrictive cardiomyopathy. Cardiac ultrasound, cardiac catheterization or magnetic resonance imaging can visualize these changes in advanced disease, but endomyocardial biopsy is the most definitive approach to establish the diagnosis.7

Recently, Cools et al. discovered the FIP1L1-PDGFRA fusion gene as a recurrent molecular abnormality in nine of 17 patients with IHES or CEL. This novel fusion gene results from a cryptic 4q12 interstitial deletion involving an 800 kb region between FIP1L1 and PDGFRA. It encodes a constitutively active FIP1L1-PDGFRA fusion protein, the tyrosine kinase activity of which is strongly inhibited by imatinib (STI-571) in vitro as well as in vivo.8,9 Taken together, these findings suggest that FIP1L1-PDGFRA (+) CEL represents a distinct clinicobiological entity of which the malignant behavior is determined by the FIP1L1-PDGFRA fusion protein. They further suggest that FIP1L1-PDGFRA (+) CEL may account for at least some of the responses of chronic hypereosinophilias to imatinib as reported recently.10,11,12 Yet, responses to imatinib have also been observed in some FIP1L1-PDGFRA (−) cases, the molecular basis of which remains unknown.8

As the clinical presentation and course of FIP1L1-PDGFRA (+) CEL is presently unknown, we have retrospectively studied 17 cases of IHES or CEL, and examined the presence of FIP1L1-PDGFRA mRNA by RT-PCR and the status of the 4q12 region by interphase fluorescence in situ hybridization (FISH). The clinical, biochemical and molecular aspects of FIP1L1-PDGFRA (+) CEL were studied and compared with FIP1L1-PDGFRA (−) CEL and IHES. Finally, we were able to monitor the hematological and molecular responses to imatinib in four surviving cases with FIP1L1-PDGFRA (+) CEL.

Materials and methods

Patients' selection

We selected 17 cases from Belgian hematology units, which fulfilled the WHO criteria for IHES/CEL,4 and for whom archived bone marrow or peripheral blood were available (Table 1). Secondary hypereosinophilias and cases with clonal T-cell populations were excluded. Case 14 was also included based on the clinical picture and evolution, although he survived only 2 months after diagnosis. As recommended by the WHO, CEL refers to cases with clonal disease, as indicated by increased blast percentages in blood (>2%) or bone marrow (5–19%), by cytogenetic markers or by nonrandom X-chromosome inactivation.4 Demonstration of the FIP1L1-PDGFRA fusion is also considered as evidence of CEL,8 while IHES is used for the remaining cases. The molecular data of patients 1, 5, 6, 9, 10 have been previously published along with a clinical summary.8 Case 16 has also been published previously.13

Table 1 Diagnostic findings

Cytogenetic analysis was performed after unstimulated short-term culture of bone marrow using standard methods. In all, 10 R-banded metaphase spreads were analyzed per case. Results are reported according to the International System for Human Cytogenetic Nomenclature (1995).14


RT-PCR detection of the FIP1L1-PDGFRA fusion was performed by nested RT-PCR using, respectively, the primers pairs FIP1L1fo: IndexTermACCTGGTGCTGATCTTTCTGA; PDGFRAro: IndexTermGCTCCCAGCAAGTTTACAATG and FIP1L1fi: IndexTermAAAGAGGATACGAATGGGACTTG; PDGFRAri: IndexTermGCCCCAGGTGAGTCATTATCT. RNA was extracted from EDTA-treated blood samples or bone marrow using Trizol LS and further purified using RNeasy columns (Qiagen) as described by the manufacturer. Total RNA (1 μg) was reverse transcribed using random hexamer primers and M-MLV reverse transcriptase from Invitrogen in a volume of 20 μl. cDNA (1 μl) was used in a 50 μl PCR. PCR fragments were detected in ethidium bromide-stained 1.5% agarose gels.

Real-time quantitative PCR (RQ-PCR)

As different fusion mRNAs occur in different patients, with splice variants in each patient sample, we designed one reverse primer IndexTerm(AACCACCTTCCCAAACGCTC) in exon 13 of PDGFRA, to be combined with forward primers developed for, respectively, the FIP1L1 exon 8/PDGFRA fusion IndexTerm(CTTCCACCGAGCAGAATGGTC, case 8), FIP1L1 exon 9 (IndexTermCGGGCAAATGAGAACAGCAA, case 6) and FIP1L1 exon 10 (IndexTermTCTTCCACCTCCTCCGACTGTC, case 7). By choosing a forward primer in the FIP1L1 exon proximal to the breakpoint, all relevant splice variants present in one sample are amplified and detected. Each primer set was validated using PCR products of the respective fusions. Total RNA (1 μg), isolated from blood, was reverse transcribed as described for the RT-PCR assay, 25% of the yield was then used for the RQ-PCR reaction in an ABI Prism 7000 Sequence Detection machine. In all, 40 cycles of annealing/extension for 1 min at 64°C were performed. PCR products were detected using SybrGreen. RQ-PCR of UBC was used as a reference to correct for sample quality15 (forward primer: ATIndexTermTTGGGTCGCGGTTCTTG; reverse primer: IndexTermTGCCTTGACATTCTCGATGGT). To determine the sensitivity of the RQ-PCR assay, EOL-1 cells carrying the FIP1L1-PDGFRA fusion16 were used. EOL-1 cells were mixed with EBV-transformed lymphoblasts at different ratios down to 0.01%. RNA was isolated and used for RQ-PCR as described for the patient samples. The presence of 0.01% EOL-1 cells resulted in a Ct value of 34.9, whereas the negative control had a Ct value of 39.4. The RQ-PCR assay thus detects less than 1 FIP1L1-PDGFRA (+) cell in 10 000. At diagnosis, 0.2–5 molecules of FIP1L1-PDGFRA fusion cDNA were detected per 1000 UBC cDNA molecules in the patient samples tested.

Fluorescence in situ hybridization

FISH analysis was performed on cytogenetic specimens of bone marrow stored at −20°C. The status of the 4q12 region was analyzed using BAC clones 120K16 (mapped centromeric of FIP1L1), 3H20 (mapped between FIP1L1 and PDGFRA) and 24O10 (mapped telomeric of PDGFRA), obtained from the RPCI11 Roswell Park Cancer Institute library ( DNA probes were directly labelled with either Spectrum Orange or Spectrum Green (Vysis, Bergisch-Gladbach, Germany) or DEAC (NEN, Zaventem, Belgium) fluorochromes conjugated with dUTP as described elsewhere.17 FISH data were collected on a Leica DMRB (Leica, Wetzlar, Germany) fluorescence microscope equipped with a cooled black and white charged-couple device camera (Photometrics, Tuscon, AZ, USA) run by Quips SmartCaptureTM FISH Imaging Software (Vysis, Bergisch-Gladbach, Germany). At least 200 interphase nuclei were scored in a blind manner for the presence or absence of the 3H20 signal, which is lost in cases with the FIP1L1-PDGFRA fusion.

Statistical analysis

Statistical analysis was performed with GraphPad Prism software, version 3.02 (GraphPad Software Inc., San Diego, CA USA) or with StatXact, version 3.0.1. Survival as of April 1, 2003 was analyzed by the log-rank test, and the results are represented as Kaplan–Meier plots.


Molecular detection and characterisation of FIP1L1-PDGFRA fusions

In all, 17 cases with IHES or CEL were selected for this study (Table 1). RT-PCR of FIP1L1-PDGFRA mRNA in five of these cases (cases 1, 5, 6, 9, 10) was presented previously, and included three FIP1L1-PDGFRA (+) cases.8 RT-PCR was performed on bone marrow or peripheral blood of the 12 remaining cases, and identified five new cases of FIP1L1-PDGFRA (+) CEL. The structure of the fusion mRNAs and of the genomic breakpoints of the latter five positive cases was determined by direct sequencing of the cloned RT-PCR products or genomic breakpoints (Figure 1a–b). As described previously, all breakpoints occurred in an FIP1L1 intron and in exon 12 of PDGFRA. Two types of fusion cDNAs could be distinguished: type I fusions result from the use of a cryptic splice acceptor site in an FIP1L1 intron resulting in an FIP1L1-PDGFRA fusion exon (cases 3, 4, 7, 8), and type II fusions in which the fusion mRNA is generated by the use of cryptic splice acceptor signals in exon 12 of PDGFRA (case 2). The breakpoints in FIP1L1 are variable and spread from introns 7 to 10. The PDGFRA breakpoints are also variable, but limited to exon 12 (Figure 1). In most patients, alternative splicing events of exons upstream of the one fused to PDGFRA occur and result in different fusion mRNAs.

Figure 1

Molecular characterization of the FIP1L1-PDGFRA fusions. (a) Structure of the different fusions. FIP1L1 sequences are indicated in bold print, dashed line or solid black box, PDGFRA sequences in normal print, solid line or white box. Coding sequences are in capitals, intron sequences in lowercase, splice donor and acceptor sites are italic. Genomic breakpoints are indicated by a black triangle. (b) Breakpoint analysis. The location of the genomic breakpoints in exon 12 of PDGFRA is indicated by triangles. Genomic breakpoints in FIP1L1 are indicated with arrows. Cases described in this work are shown below (new cases in solid black arrow, patients reported in Cools et al8 are in solid gray arrow). Patients reported in Klion et al19 are in dashed arrows (case numbers with asterisk).

Interphase FISH of 4q12 and RT-PCR are equivalent diagnostic tools for the detection of FIP1L1-PDGFRA (+) CEL

The variability of the breakpoints in the FIP1L1 gene raised the question as to whether cases with variant FIP1L1 breakpoints might escape detection by RT-PCR. Therefore, in 10 cases with available cytogenetic specimens, we examined the 4q12 region with an interphase FISH assay, using two clones (3H20 mapped between PDGFRA and FIP1L1 (Spectrum Green labelled), and 24O1O mapped telomeric to PDGFRA (Spectrum Orange)) (Figure 2a, Table 1). In cases 2–5 and 8, the loss of one green 3H20 signal, indicative of the interstitial 4q12 deletion, was found in 41–91% of informative interphase nuclei. It should be noted that the high autofluorescence of the eosinophilic granules precluded evaluation of a proportion of mature eosinophils. In cases 13–17, that were FIP1L1-PDGFRA (−), the loss of one 3H20 signal was observed in only 0.5–1.5% of the nuclei, defining the background of the assay. In cases 4 and 8, with normal karyotype and available metaphase spreads, a three-color FISH assay with the additional 120K16 clone labelled with DEAC (shown as yellow, Figure 2c and d) was performed. In both cases, the loss of one 3H20 signal had occurred on a seemingly normal looking chromosome 4 (example in Figure 2e). Thus in all cases examined, interphase FISH and RT-PCR yielded equivalent results, indicating that both assays are valid diagnostic tools.

Figure 2

FISH detection of the 4q12 deletion associated with the FIP1L1-PDGFRA fusion. (a) A genomic map of the 4q12 region, with relevant genes and selected FISH probes is drawn to scale. Examples are shown of three-color FISH analysis performed on a control sample (b), patient 4 (c) and patient 8 (d and e), showing the loss of one 3H20 (green) signal in (c–e) (arrow). Arrow in (e) indicates a seemingly normal looking chromosome 4 with the cryptic del(4)(q12).

FIP1L1-PDGFRA fusion gene is the sole marker of clonality in the majority of FIP1L1-PDGFRA fusion (+) CEL

In the FIP1L1-PDGFRA (+) group, only patient 1 had an elevated marrow blast percentage at diagnosis, as additional evidence of clonality. In the second year after diagnosis, she acquired a jumping 1q12 translocation and nonrandom X-chromosome inactivation (as assessed by HUMARA assay) was shown in peripheral blood eosinophils and neutrophils, but not T-lymphocytes or monocytes (data not shown). Patient 8 developed an elevated peripheral blastosis at month 26 after diagnosis. Therefore, in this series, the FIP1L1-PDGFRA fusion gene was the unique marker of clonality in six out of eight cases (Table 1).

Clinical and laboratory observations at presentation

The male/female ratio in this series was 13/4 with a median age at diagnosis of 47.7 years (range 26.5–62.4 years). Based on molecular, cytogenetic and cytological (central or peripheral blastosis) data, the cases studied are classified as FIP1L1-PDGFRA (+) CEL (cases 1–8), FIP1L1-PDGFRA (−) CEL (cases 15–17) and IHES (cases 9–14). On first presentation, all patients with CEL and five of six patients with IHES were symptomatic, with fatigue as most commonly reported complaint. Palpable splenomegaly was common in FIP1L1-PDGFRA (+) CEL (5/8 cases), in FIP1L1-PDGFRA (−) CEL (3/3 cases) but not in IHES (only 1/6 cases) (Tables 1 and 2). Therefore splenomegaly was significantly more frequent in CEL, as reported previously (P=0.050 in Fisher's test). In addition, the vitamin B12 serum level was strongly increased outside the routinely measured range (>1800–2000 ng/l) in all evaluable cases with FIP1L1-PDGFRA (+) or FIP1L1-PDGFRA (−) CEL, but was normal in IHES (P=0.0001 by Fisher's test). Hemoglobin level, circulating absolute eosinophil, leukocyte and platelet counts were not significantly different between the groups with FIP1L1-PDGFRA (+) CEL, FIP1L1-PDGFRA (−) CEL or IHES.

Table 2 Clinical, hematological and laboratory variables at diagnosis in FIP1L1-PDGFRA (+) CEL, FIP1L1-PDGFRA (−) CEL and IHES

Serial bone marrow biopsies available for cases 1, 3, 8 and 9 were reviewed. In FIP1L1-PDGFRA (+) CEL, they revealed a markedly hyperplastic and proliferative myelopoiesis, suggestive of a myeloproliferive disorder-type CML, and with scattered spindle-shaped C-KIT and tryptase-positive mast cells. Moreover, these findings are clearly distinct from case 9 with IHES, with a normocellular marrow, and lower numbers of C-KIT and tryptase-positive mast cells (data not shown), as well as from systemic mastocytosis with eosinophilia where focal aggregates of tryptase-positive mast cells are found.18

Organ involvement at diagnosis and during evolution

Two patients with FIP1L1-PDGFRA (+) CEL (cases 5 and 7) had clinically important and symptomatic eosinophilic endomyocardial disease at first presentation, as suggested by findings by noninvasive cardiac imaging studies. Two other patients (case 2 and 8) had a normal echocardiographic examination at diagnosis but developed eosinophilic endomyocardial disease after 4 and 18 months, respectively. Of note, in patient 2, cardiac ultrasound evidenced left ventricular dysfunction and apical hypertrophy 4 months after diagnosis, but revealed a dilated cardiomyopathy only 2 months later, a finding confirmed in several later examinations. Finally, patient 6 was also found to have a dilated left ventricular cavity at diagnosis, in the absence of an obvious other explanation. We have not identified cases with eosinophilic heart disease among FIP1L1-PDGFRA (−) CEL. In the group classified as IHES, two of six (cases 9 and 14) developed eosinophilic endomyocardial disease. Case 9 was symptomatic from first presentation onwards, and had biopsy-proven eosinophilic endomyocardial disease. Patient 14, despite a normal cardiac ultrasound study at diagnosis, developed an infiltrative cardiomyopathy with altered diastolic relaxation within the next 2 months. He died a sudden unwitnessed, presumably cardiac death while admitted for uncontrolled eosinophil leukocytosis, but autopsy was not performed. Thrombotic events developed only in the FIP1L1-PDGFRA (+) group (cases 5, 7 and 8, Table 3).

Table 3 Evolution of FIP1L1-PDGRA (+) CEL, FIP1L1-PDGRA (−) CEL and IHES

Hematological evolution, response to therapy and survival of FIP1L1-PDGFRA (+) CEL, FIP1L1-PDGFRA (−) CEL and IHES

In the group with FIP1L1-PDGFRA (+) CEL, patients 2, 3 and 8 developed marked peripheral blood leukocytosis above 50 × 109/l with marked eosinophilia early after diagnosis (Table 3). This was precipitated by a diagnostic splenectomy in patient 2, while in patients 3 and 8 this occurred despite medical therapy. In patients 1 and 4, high peripheral blood leukocyte and eosinophil counts developed only in the preterminal disease phase. Of note, increasing leukocytosis and eosinophilia were accompanied by increased peripheral or central blastosis only in patients 1 and 8. Before the availability of imatinib, medical treatment consisted of corticosteroids, hydroxyurea and interferon-α, either alone or in combination: these therapies induced only partial and/or transient clinical improvement, often at the expense of anemia and thrombocytopenia. Among the agents used, hydroxyurea was the most efficacious, while steroids, used mostly in conjunction with other therapies, had limited effect, and interferon-α had to be discontinued in all cases used either because of intolerance or lack of effect. Patient 3 underwent a related HLA-matched bone marrow transplantation on day 482 after diagnosis, and remains alive in remission. Of note, he no longer has detectable FIP1L1-PDGFRA mRNA in his peripheral blood, but retains a mild and stable eosinophilia (0.6–2.0 × 109/l), maybe due to limited chronic GVHD. With the exception of patient 11 who is untreated, patients with IHES or FIP1L1-PDGFRA (−) CEL were treated with a variety of agents, including corticosteroids, hydroxyurea, interferon-α, cyclosporin A, antileukemic chemotherapy or combinations. Responses were variable and ranged from refractoriness (cases 14 and 17), over partial hematological responses (cases 9, 10, 12 and 13) to complete and sustained hematological responses to single-agent interferon-α (cases 15 and 16)13 (Table 3). In the group with FIP1L1-PDGFRA (+) CEL, after a median follow-up of 1406 days, four of eight patients have died. All these deaths were disease related, either from complications of cardiac failure (patient 5), uncontrolled disease activity with cardiopulmonary failure (patients 2 and 4) or evolution to acute myeloid leukemia (patient 1) (Figure 3a and b). In the IHES group, patient 14 had a steeply increasing leukocyte and eosinophil count within months after diagnosis and died a sudden, presumably cardiac death. Of the FIP1L1-PDGFRA (−) CEL, patient 15 eventually died from complicated surgery in the setting of a neurological disease not known to be associated with eosinophilia. Patient 17 died from infectious transplant-related mortality after a matched unrelated bone marrow transplantation 4.5 years after diagnosis.

Figure 3

Survival of FIP1L1-PDGFRA (+) CEL, FIP1L1-PDGFRA (−) CEL and IHES. (a) Shows overall survival. (b) Shows specific survival (excluding nondisease-related deaths: cases 3 and 17 are censored at days 482 and 1641 (day 0 of allogeneic BMT) respectively, patient 15 at day 2733 (time of death after gastrostomy surgery).

Imatinib induces complete hematological and partial or complete molecular responses in FIP1L1-PDGFRA (+) CEL

Since the availability of imatinib, four FIP1L1-PDGFRA (+) patients (cases 5–8) have been treated at an initial dose of 100 mg/day, followed by a tapered maintenance dose. In all of the treated patients, including the female patient 8, imatinib induced a rapid and complete hematological response with normalization of the peripheral eosinophil count (data not shown). Nevertheless, no clear improvement of cardiac failure was observed so far in patients 5, 7 and 8, and patient 5 died from cardiac failure a few weeks later. The presence of FIP1L1-PDGFRA mRNA was analyzed in the blood of the three surviving patients under treatment: the fusion became undetectable by nested PCR in patients 6 and 7 and remained so during 4 months of follow-up. The third sample from the female patient 8 remained positive by nested RT-PCR. To evaluate the load of FIP1L1-PDGFRA fusion mRNA in this patient a RQ-PCR was developed. Samples obtained before and after treatment were available for patients 6, 7 and 8. UBC was used as a reference gene for normalization of the results. Consistent with results of nested RT-PCR, the fusion became undetectable by RQ-PCR in two out of three patients (patients 6 and 7, threshold values for detetion of a product >40 cycles). In the female (patient 8), the fusion RNA remained detectable by RQ-PCR with approximately a 10-fold reduction of the FIP1L1-PDGFRA/UBC ratio (Figure 4). So far, only patient 9 in the IHES group has been treated with imatinib, which was rapidly abandoned for intolerance and without evidence of response.

Figure 4

RQ-PCR analysis of the FIP1L1-PDGFRA fusion during imatinib treatment. RQ-PCR of the different FIP1L1-PDGFRA fusions was performed and normalized for the amount of UBC cDNA present in the sample. The FIP1L1-PDGFRA/UBC ratio was then normalized against the values observed at diagnosis. FIP1L1-PDGFRA cDNA load normalized against the load at day 0 (y-axis) is shown as a function of time (x-axis, days of imatinib treatment).


Demonstration of the FIP1L1-PDGFRA fusion in patients with sustained hypereosinophilia provides evidence of clonality and thus supports the diagnosis of CEL.8 In this series of patients fulfilling WHO criteria of IHES/CEL, we present eight patients with the FIP1L1-PDGFRA fusion gene, and have compared their phenotype with IHES and FIP1L1-PDGFRA (−) CEL. Our data indicate that FIP1L1-PDGFRA (+) CEL represents a homogeneous clinical entity with the hallmarks of the hypereosinophilic syndrome,3,7 including a high probability of cardiac complications and thrombosis, a predilection for middle-aged males, a poor response to conventional medical treatment and a high probability of disease-related death. Four of eight FIP1L1-PDGFRA (+) cases developed cardiac failure early in their disease course, with a typical restrictive echocardiographic pattern or post-mortem findings of eosinophilic endomyocarditis. Two cases had dilated cardiomyopathy at some point in their disease, lacking other plausible explanations. Thus, these findings may suggest that cardiac involvement in hypereosinophilia may also encompass dilated cardiomyopathy. Finally, the data also suggest that under conventional medical therapy of the preimatinib era, the diagnosis of FIP1L1-PDGFRA (+) CEL implies a worse prognosis than IHES (survival range in this series 30–78 months after diagnosis).

From our series, it also emerges that a strongly elevated vitamin B12 serum level is surrogate marker strictly associated with CEL, either FIP1L1-PDGFRA (+) or FIP1L1-PDGFRA (−). Elevated vitamin B12 levels as well as splenomegaly were previously reported in approximately half of IHES/CEL patients in series published by French and NIH investigators.2,5,6,7 More recently, in a subset of 15 patients with idiopathic IHES, excluding cases with cytogenetic markers, Klion et al19 reported elevated serum tryptase levels in nine patients. The latter nine cases with elevated serum tryptase also had a strongly elevated serum vitamin B12 level, seven of nine had splenomegaly, and the FIP11L-PDGFRA fusion was detected by RT-PCR in five of five tested cases. In contrast, four IHES cases with normal serum tryptase in the Klion series tested negative for the FIP11L-PDGFRA fusion.19 Although the relationship between elevated serum tryptase and FIP1L1-PDGFRA (+), CEL FIP1L1-PDGFRA (−) CEL and IHES could not be investigated in our series, the hyperplastic myelopoiesis and pattern of C-KIT and tryptase-positive mast cells that we found in bone marrow biopsies of FIP1L-PDGFRA (+) patients is extremely similar to what Klion et al19 found in cases with elevated serum tryptase, and clearly distinct from the normocellularity observed in a case with IHES. Taken together, our and their data collectively suggest that the FIP1L1-PDGFRA fusion gene, a strongly elevated serum vitamin B12 level, or an elevated serum tryptase level define strongly overlapping groups of patients with myeloproliferative hypereosinophilias. Consistent with the myeloproliferative nature of these hypereosinophilias, X-chromosome inactivation assays in patient 1 showed clonality in peripheral blood eosinophil and neutrophil fractions but not monocytes or T-lymphocytes (data not shown). In addition, the FIP1L1-PDGFRA fusion gene could not be demonstrated in a patient not included in this series with hypereosinophilia and a clonal T-cell population.20 The diagnostic value of bone marrow biopsy in the work up of chronic idiopathic eosinophilias also requires further study.

Of FIP1L1-PDGFRA (+) cases, four died from a disease-related death, one case was rescued by allogeneic bone marrow transplantation and three recently diagnosed cases are in hematological remission under imatinib. Interestingly, in the series by Klion et al.19 three of nine serum tryptase-positive cases died a disease-related death. We have also observed one disease-related death and two cases with eosinophilic endomyocarditis among FIP1L1-PDGFRA (−) IHES, contrary to Klion et al,19 who found no disease-related deaths or cardiac complications in IHES with normal tryptase. The reasons for these discrepancies are unclear at the moment.

Recently, the FIP1L1-PDGFRA fusion has also been reported in systemic mast cell disease with eosinophilia.18 The latter disease has clinical features that are clearly distinct from the presentation of FIP1L1-PDGFRA (+) CEL. In addition, histopathological examination of available bone marrow biopsies from three of our cases with the FIP1L1-PDGFRA fusion failed to reveal the multifocal dense infiltrates of atypical mast cells that are typical of systemic mast cell disease with eosinophilia.21 Yet, scattered dysplastic spindle-shaped mast cells with positivity for CKIT and tryptase were present in these bone marrow biopsies. Therefore, assessment of the other serological and biochemical features of systemic mast cell disease could have allowed to further support our judgment that FIP1L1-PDGFRA (+) CEL is a disease different from FIP1L1-PDGFRA (+) systemic mast cell disease with eosinophilia.18,21 Unfortunately, as patient materials are not available, these further analyses cannot be done.

FIP1L1-PDGFRA (+) CEL was most often not associated with other detectable cytogenetic abnormalities or elevated central or peripheral blastosis, but was subject to clonal evolution and leukemic transformation in some cases. Therefore, it seems reasonable to investigate the presence of this fusion gene in all cases of sustained hypereosinophilia without aberrations in conventional cytogenetic analysis, or with cytogenetic aberrations not specifically associated with hypereosinophilia. Moreover, as the FIP1L1-PDGFRA is by itself an evidence of clonality, it seems appropriate to consider all FIP1L1-PDGFRA (+) hypereosinophilias as CEL, regardless of the presence or absence of other clonality markers.

FISH experiments in RT-PCR-negative patients have not revealed any alternative 4q12 rearrangements, involving variant fusion partners for FIP1L1 or PDGFRA, which might be missed by RT-PCR. Thus, both approaches seem equivalent in a diagnostic setting. However, eosinophil autofluorescence in interphase FISH can make it difficult to evaluate the exact percentage of FIP1L1-PDGFRA (+) cells. Therefore, FISH might be less suitable for follow-up of FIP1L1-PDGFRA (+) CEL under therapy. In this series, we have treated four FIP1L1-PDGFRA (+) patients with imatinib, and observed complete and sustained clinical and hematological responses in all of them, as previously reported for five other cases with FIP1L1-PDGFRA (+) CEL.8 In addition, Klion et al19 have reported complete responses to imatinib in patients with hypereosinophilia and elevated serum tryptase, four of whom turned out to be FIP1L1-PDGFRA (+). While sustained hematological responses to imatinib were until now observed only in males with FIP1L1-PDGFRA (+) CEL or IHES,10,11,12 we document here the first sustained hematological response to imatinib in a female. Yet, she achieved only a limited reduction in fusion gene mRNA in contrast to the complete molecular response in the two male cases. Thus, our data also suggest that RT- or RQ-PCR can be used to monitor residual disease in FIP1L1-PDGFRA (+) CEL, and that they can be helpful in adjusting the dosage of imatinib. The gratifying clinical, hematological and molecular responses with imatinib in this setting are unprecedented, but more cases and longer follow-up will be required to resolve definitively whether imatinib can induce long-term molecular remission in this disease.


  1. 1

    Bain BJ . Eosinophilic leukaemias and the idiopathic hypereosinophilic syndrome. Br J Haematol 1996; 95: 2–9.

    CAS  PubMed  Google Scholar 

  2. 2

    Bain BJ . Hypereosinophilia. Curr Opin Hematol 2000; 7: 21–25.

    CAS  Article  Google Scholar 

  3. 3

    Chusid MJ, Dale DC, West BC, Wolff SM . The hypereosinophilic syndrome: analysis of fourteen cases with review of the literature. Medicine (Baltimore) 1975; 54: 1–27.

    CAS  Article  Google Scholar 

  4. 4

    Bain B, Pierre R, Imbert M, Vardiman JW, Brunning RD, Flandrin G . Chronic eosinophilic leukemia and the hypereosinophilic syndrome. In: Jaffe ES, Harris NL, Vardiman JW (eds) WHO Classification of Tumours: Pathology and Genetics. Tumors of Haematopoietic and Lymphoid tissues. Lyon: IARC Press, 2001, pp 29–31.

    Google Scholar 

  5. 5

    Zittoun J, Farcet JP, Marquet J, Sultan C, Zittoun R . Cobalamin (vitamin B12) and B12 binding proteins in hypereosinophilic syndromes and secondary eosinophilia. Blood 1984; 63: 779–783.

    CAS  PubMed  Google Scholar 

  6. 6

    Fauci AS, Harley JB, Roberts WC, Ferrans VJ, Gralnick HR, Bjornson BH . NIH conference. The idiopathic hypereosinophilic syndrome. Clinical, pathophysiologic, and therapeutic considerations. Ann Intern Med 1982; 97: 78–92.

    CAS  Article  Google Scholar 

  7. 7

    Weller P, Bubley G . The idiopathic hypereosinophilic syndrome. Blood 1994; 83: 2759–2779.

    CAS  Google Scholar 

  8. 8

    Cools J, DeAngelo DJ, Gotlib J, Stover EH, Legare RD, Cortes J et al. A tyrosine kinase created by fusion of the PDGFRA and FIP1L1 genes as a therapeutic target of imatinib in idiopathic hypereosinophilic syndrome. N Engl J Med 2003; 348: 1201–1214.

    CAS  Article  Google Scholar 

  9. 9

    Cools J, Stover EH, Boulton CL, Gotlib J, Legare RD, Amaral SM et al. PKC412 overcomes resistance to imatinib in a murine model of FIP1L1-PDGFRalpha-induced myeloproliferative disease. PG-459-69. Cancer Cell 2003; 5: 459–469.

    Article  Google Scholar 

  10. 10

    Gleich GJ, Leiferman KM, Pardanani A, Tefferi A, Butterfield JH . Treatment of hypereosinophilic syndrome with imatinib mesilate. Lancet 2002; 359: 1577–1578.

    CAS  Article  Google Scholar 

  11. 11

    Pardanani A, Reeder T, Porrata LF, Li C-Y, Tazelaar HD, Baxter EJ et al. Imatinib therapy for hypereosinophilic syndrome and other eosinophilic disorders. Blood 2003; 101: 3391–3397.

    CAS  Article  Google Scholar 

  12. 12

    Cortes J, Ault P, Koller C, Thomas D, Ferrajoli A, Wierda W et al. Efficacy of imatinib mesylate in the treatment of idiopathic hypereosinophilic syndrome. Blood 2003; 101: 4714–4716.

    CAS  Article  Google Scholar 

  13. 13

    Malbrain ML, Van den Bergh H, Zachee P . Further evidence for the clonal nature of the idiopathic hypereosinophilic syndrome: complete haematological and cytogenetic remission induced by interferon-alpha in a case with a unique chromosomal abnormality. Br J Haematol 1996; 92: 176–183.

    CAS  Article  Google Scholar 

  14. 14

    Mitelman F, ISCN. An International System for Human Cytogenetic Nomenclature. Basel: S Karger, 1995.

    Google Scholar 

  15. 15

    Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol 2002; 3, RESEARCH0034.1-0034.11.

  16. 16

    Cools J, Quentmeier H, Huntly BJ, Marynen P, Griffin JD, Drexler HG et al. The EOL-1 cell line as an in vitro model for the study of FIP1L1-PDGFRA positive chronic eosinophilic leukemia. Blood 2003, Nov 20 [Epub ahead of print], in press.

  17. 17

    Martin-Subero JI, Harder L, Gesk S, Schlegelberger B, Grote W, Martinez-Climent JA et al. Interphase FISH assays for the detection of translocations with breakpoints in immunoglobulin light chain loci. Int J Cancer 2002; 98: 470–474.

    CAS  Article  Google Scholar 

  18. 18

    Pardanani A, Ketterling RP, Brockman SR, Flynn HC, Paternoster SF, Shearer BM et al. CHIC2 deletion, a surrogate for FIP1L1-PDGFRA fusion, occurs in systemic mastocytosis associated with eosinophilia and predicts response to imatinib mesylate therapy. Blood 2003; 102: 3093–3096.

    CAS  Article  Google Scholar 

  19. 19

    Klion AD, Noel P, Akin C, Law MA, Gilliland DG, Cools J et al. Elevated serum tryptase levels identify a subset of patients with a myeloproliferative variant of idiopathic hypereosinophilic syndrome associated with tissue fibrosis, poor prognosis, and imatinib responsiveness. Blood 2003; 101: 4660–4666.

    CAS  Article  Google Scholar 

  20. 20

    Roufosse FE, Goldman M, Cogan E . Hypereosinophilic syndrome. N Engl J Med 2003; 348: 2687.

    Article  Google Scholar 

  21. 21

    Valent P, Horny HP, Li CY, Longley BJ, Metcalfe DD, Parwaresch RM et al. Mastocytosis. In: Jaffe ES, Harris NL, Vardiman JW (eds) WHO Classification of Tumours: Pathology and Genetics. Tumors of Haematopoietic and Lymphoid tissues. Lyon: IARC Press, 2001, pp 292–302.

    Google Scholar 

Download references


We thank Dr C Peeters-De Wolf for critical reading and helpful discussions, Ursula Pluys for technical assistance with FISH, Drs H Van Keer, J Billiet and M Pilet for retrieving laboratory data and G Verbeke for statistical advice. This paper presents research results of the Belgian Programme of Interuniversity Poles of Attraction, initiated by the Belgian State, Prime Minister's Office, Science Policy Programming. The scientific responsibility is assumed by the authors. PV is a senior clinical investigator, JC a postdoctoral research fellow of Fonds voor Wetenschappelijk Onderzoek Vlaanderen. LM is partially supported by the Salus Sanguinis Foundation. This work was supported by grant G.0120.00 of the Fonds voor Wetenschappelijk Onderzoek Vlaanderen.

Author information



Corresponding authors

Correspondence to P Vandenberghe or P Marynen.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Vandenberghe, P., Wlodarska, I., Michaux, L. et al. Clinical and molecular features of FIP1L1-PDFGRA (+) chronic eosinophilic leukemias. Leukemia 18, 734–742 (2004).

Download citation


  • chronic eosinophilic leukemia
  • idiopathic hypereosinophilic syndrome
  • imatinib

Further reading


Quick links