EBF1-PDGFRB fusion in pediatric B-cell precursor acute lymphoblastic leukemia (BCP-ALL): genetic profile and clinical implications

Chromosomal abnormalities are the hallmark of B-cell precursor acute lymphoblastic leukemia (BCP-ALL) and have prognostic relevance.¹  Integration of minimal residual disease (MRD), cytogenetics, age, and white blood cell count into risk stratification for treatment of childhood BCP-ALL has contributed significantly to improved survival rates.²  Approximately 25% of patients with childhood BCP-ALL harbor none of the established chromosomal abnormalities termed “B-other.” A novel subgroup of B-other BCP-ALL has been described and is known as Philadelphia-like (Ph-like) or BCR-ABL1-like ALL.^3,4  Although these patients lack the BCR-ABL1 fusion, they have gene expression profiles and high risk of relapse similar to patients with BCR-ABL1-positive ALL. A subset of Ph-like patients harbor tyrosine kinase–activating gene fusions,⁵  notably EBF1-PDGFRB, which accounts for ∼8% of patients⁶ ; a proportion of those patients were refractory to conventional therapy but they achieved a complete response when treated with the tyrosine kinase inhibitor imatinib.^6-8  Here we present genetic and clinical data from 15 EBF1-PDGFRB-positive patients treated in childhood ALL treatment trials in the United Kingdom.

All patients had BCP-ALL and were registered in treatment trials in the United Kingdom (supplemental Figures 1 and 2, available on the Blood Web site): Medical Research Council ALL 97/99 (1997-2002) (age 1-18 years), UKALL 2003 (2003-2011) (age 1-24 years), UKALL 2011 (2012-present) (age 1-24 years) (www.isrctn.com/ISRCTN64515327), and relapse trial ALLR3 (2003-2013), with ethical approval and consent in accordance with the Declaration of Helsinki.^9-11  Demographic, clinical, and treatment details were collected by the Clinical Trial Service Unit, Oxford, United Kingdom. Cytogenetic analysis was performed in regional cytogenetics laboratories and collated by the Leukaemia Research Cytogenetics Group. EBF1-PDGFRB was determined by fluorescence in situ hybridization (FISH) using a commercial PDGFRB break-apart probe (Cytocell, Cambridge, United Kingdom) (supplemental Figure 3). Involvement of EBF1 was confirmed by using in-house break-apart probes (supplemental Figure 3). Copy number changes were identified by multiplex ligation-dependent probe amplification (MLPA) (SALSA MLPA kit P335 IKZF1, MRC Holland, Amsterdam, The Netherlands) (n = 11) (supplemental Figure 4)^12,13  and SNP 6.0 (n = 9) (AROS Applied Biotechnology, Aarhus, Denmark, and Genotyping Console, Affymetrix, Santa Clara, CA) mapped to human reference sequence GRCh37. The presence of EBF1-PDGFRB fusion transcripts was validated by reverse transcription polymerase chain reaction (RT-PCR) (n = 9) (supplemental Figure 5).

Patient data are provided in Table 1 and supplemental Table 1 (patient origin provided in supplemental Figures 1 and 2). For 11 of 15 patients (patients 1, 2, 4, and 7-14), FISH for EBF1 and PDGFRB showed signal patterns consistent with deletion of 5q33, with break points within EBF1 and PDGFRB (Figure 1; supplemental Figure 3), as previously reported.^5,7,8  Deletion of EBF1 exon 16 was seen by MLPA in 8 patients and was tested and confirmed by SNP 6.0 in 6 of them (supplemental Figure 4; supplemental Table 1). In patients who had available RNA, EBF1-PDGFRB transcripts were confirmed by RT-PCR (n = 9; patients 6, 7, and 9-15). EBF1 exon 15 was fused to PDFGRB exon 11 and was confirmed by Sanger sequencing (n = 4), except for patient 13 who had an alternative break point (supplemental Figure 5).

Two patient samples showed signal patterns consistent with balanced rearrangements by both PDGFRB and EBF1 FISH (patients 5 and 6) (Figure 1; supplemental Figure 3). Expression of the EBF1-PDGFRB transcript was confirmed by RT-PCR in patient 6 (supplemental Figure 5). This patient showed cytogenetic evidence of additional material on the long arm of chromosome 5 (5q). No copy number abnormalities of chromosome 5 were detected by SNP 6.0 (data not shown); thus, the karyotype was suggestive of a translocation t(5;5)(q33.1;q33.3) (Figure 1). However, poor metaphase quality in this patient and in patient 5 precluded confirmation of this subtle abnormality.

Despite showing an apparently balanced rearrangement of PDGFRB and EBF1 by FISH (supplemental Figure 3), SNP 6.0 analysis of a sample from patient 3 revealed a series of nonconsecutive deletions along 5q, indicating that complex rearrangements may have resulted in the disruption of these genes (Figure 1).

In patient 15, the EBF1-PDGFRB transcript was detected by RT-PCR (supplemental Figure 5), although FISH showed a balanced abnormal signal pattern for PDGFRB and an unbalanced abnormal pattern for EBF1 (supplemental Figure 3). Because of a lack of material, the karyotypic nature of the rearrangement was not determined.

Among the genes tested by MLPA, those often deleted in BCP-ALL were also deleted in this cohort^3-6,12 : PAX5 (n = 5), IKZF1 (n = 3), and CDKN2A/B (n = 3) (supplemental Table 1).

In contrast to ALL overall, there was predominance of females compared to males (11:4), median age was 12 years (range, 4-18 years; >10 years [n = 9]), and median white blood cell count was 48.8 × 10⁹/L (range 3.4-345 × 10⁹/L; >50 × 10⁹/L [n = 9]). From an unselected cohort of 287 B-other patients from UKALL 2003 (supplemental Figure 1), 2.7% (n = 8) were EBF1-PDGFRB positive, indicating an incidence of ∼0.5% in BCP-ALL overall, on the basis of the mutual exclusivity of this translocation and its occurrence being restricted to the B-other group.

Among 13 patients from the early trials (ALL 97/99, n = 3; UKALL 2003, n = 10), 12 achieved complete remission (CR), whereas 1 did not achieve remission and died after 2 months. Ten remitters had slow response (all from UKALL 2003) and were positive for MRD at the end of induction, including 8 with extremely high MRD levels (>10%). Consequently, 8 of 13 received augmented postinduction therapy (regimen C)¹⁰  and 3 had bone marrow transplants in first CR. Seven patients relapsed 18 to 59 months after diagnosis. The majority (9 of 13) remained alive at 6 to 9 years, and 4 died (nonremitter, n = 1; infection in remission, n = 1; relapse, n = 2). Recently, the Dutch Childhood Oncology Group reported 4 EBF1-PDGFRB-positive patients with 1 relapse and 3 remaining in first CR at 6 to 19 years.¹⁴  Clinical outcomes for that study and our study correlate with results from the Children’s Oncology Group, which showed that although Ph-like patients had poor initial response to treatment, their survival rate improved when they were treated with MRD-based risk-directed therapy.¹⁵ 

As a result of these findings, we screened the following population for the EBF1-PDGFRB fusion (supplemental Figure 2): patients entered onto the current UKALL 2011 trial in the United Kingdom who belonged to the B-other group and who had failure of induction therapy, did not achieve CR by day 29, or who remained MRD positive (>0.5%) at week 14. Thus far, 2 patients have tested positive for EBF1-PDGFRB: a 5-year-old female (patient 14) who was MRD positive at day 35 (30%) and an 8-year-old female (patient 15) with poor response to induction therapy. Detection of EBF1-PDGFRB in patient 14 prompted her withdrawal from the trial, and imatinib was added (off-label) to the chemotherapy regimen used in the EsPhALL trial.¹⁶  After 5 weeks of Protocol 1B consolidation therapy with imatinib, her MRD levels fell to 6% and she became MRD negative after 3 postinduction treatment blocks (HR1, HR2, and HR3). Given her persistent MRD, she subsequently received a matched unrelated bone marrow transplant, but unfortunately she died 6 months later as a result of undefined encephalopathy. Patient 15 had high levels of MRD after induction (30%). She was treated with imatinib in addition to augmented Berlin-Frankfurt-Munster consolidation¹⁰  and became MRD negative at week 14. She continues in CR at 10 months after diagnosis and is currently receiving maintenance therapy and continuous imatinib.

This study reports the largest cohort of EBF1-PDGFRB patients to date. It demonstrates the range of genetic mechanisms by which the fusion may occur. EBF1-PDGFRB fusion is associated with female sex, older age, and a varied outcome on treatment trials in the United Kingdom. Although these patients had high levels of MRD and tendency to relapse, there was also evidence of durable remission, especially with intensive chemotherapy. Evidence from this study and others^6-8  suggests that these patients respond effectively to imatinib. It is interesting to speculate whether treatment with tyrosine kinase inhibitors may avoid the need for intensive chemotherapy to achieve a cure.

The authors thank member laboratories of the UK Cancer Cytogenetic Group for providing cytogenetic data and material. Primary childhood leukemia samples used in this study were provided by the Leukaemia and Lymphoma Research Childhood Leukaemia Cell Bank.

This work was supported by Bloodwise (formerly Leukaemia and Lymphoma Research, United Kingdom).

Contribution: C.S., A.V.M., and C.J.H. designed the study; C.S., S.L.R., A.E., J. Murray, S.R., and C.W. carried out the cytogenetic and molecular testing; C.S., L.C., C.J.H., and A.V.M analyzed and interpreted the data; J. Moppett, M.C., O.T., C.A.P., V.S., N.G., and A.V. provided clinical and follow-up data; C.J.H. and A.V.M. provided financial and administrative support; and all authors wrote and provided final approval of the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Christine J. Harrison, Leukaemia Research Cytogenetics Group, Northern Institute for Cancer Research, Newcastle University, Level 5, Sir James Spence Institute, Royal Victoria Infirmary, Newcastle-upon-Tyne NE1 4LP, United Kingdom; e-mail: christine.harrison@newcastle.ac.uk.