ETV6 and PDGFRB: a license to fuse

In 1994 the ETV6-PDGFRB (TEL-PDGFRB) fusion gene was identified by Todd Golub and Gary Gilliland in patients with chronic myelomonocytic leukemia (CMML) and t(5;12)(q33;p13).1 The implications of the molecular characterization of this translocation go far beyond the identification of the ETV6-PDGFRB oncogene. The identification of the ETV6 gene on chromosome 12p13 has led to the discovery of a large number of important fusion genes such as the ETV6-RUNX1 fusion that is present in up to 25% of childhood B-cell acute lymphocytic leukemia (B-ALL).2 In addition, a wide variety of other fusion partners of PDGFRB have also been identified. In this issue of the journal Walz and colleagues report on the identification of yet three more novel PDGFRB fusion genes in chronic myeloproliferative disorders.3 The identification of a PDGFRB fusion gene in the leukemic cells of a patient is of major importance since these patients can be treated with the small molecule kinase inhibitor imatinib (STI-571, Gleevec, Glivec).4 All patients with myeloproliferative disorders characterized by expression of a PDGFRB fusion gene respond well to imatinib therapy, with rapid and complete hematologic and molecular remissions observed in these patients. Thus, the identification of a translocation involving the PDGFRB gene in patients with myeloproliferative disorders is a clear marker that predicts response to imatinib therapy.

Translocations involving the ETV6 gene

Since the initial identification of ETV6 as a gene rearranged by the t(5;12)(q33;p13), a large number of variant translocations that also involve ETV6 have been reported (Table 1). Some of these translocations result in fusions of ETV6 to other tyrosine kinase genes such as JAK2,5 ABL1,6 NTRK3,7 and FLT3.8 The common theme here is the fusion of the homodimerization domain of ETV6 to a tyrosine kinase domain, resulting in the generation of a constitutively active kinase. These activated kinases phosphorylate themselves as well as a variety of downstream signaling proteins, which leads to the stimulation of proliferation and survival pathways. Most of these translocations are rare, but remain interesting to detect, as these patients are likely to respond to treatment with small molecule inhibitors interfering with the activity of these oncogenic tyrosine kinase genes. The fusion of ETV6 to tyrosine kinase genes is, however, not the only mechanism by which ETV6 contributes to leukemogenesis. Different parts of ETV6 can also be fused to transcription factors, such as RUNX1 (AML1),9 EVI1,10 and MN1.11 The most important translocation in this subgroup is the cryptic t(12;21)(p13;q22) that is found in 25% of childhood B-ALL cases, and results in the fusion of ETV6 to RUNX1.2 The t(12;21) is frequently associated with deletion or inactivation of the other ETV6 allele, and is believed to confer a favorable prognosis. ETV6-RUNX1 (TEL-AML1) acts as an aberrant transcription factor that is believed to be involved in the expansion of hematopoietic progenitors,12 but the exact way ETV6-RUNX1 alters the differentiation and self-renewal pathways remains unclear.

Table 1.Translocations involving the ETV6 gene.

Most of the chromosomal translocations involving ETV6 result in the generation of some kind of fusion gene, but it has not always been clear whether these fusions are the relevant oncogenic events or not.13,14 In fact, other consequences of the chromosomal translocations may also be important. In this context, it was proposed that translocations involving ETV6 could also result in a deregulated expression of nearby oncogenes, as in the case of the translocations t(4;12) and t(5;12), resulting in expression of GSH2 and IL3 respectively.15 Finally, ETV6 was also found to be mutated in acute myeloid leukemia,16 revealing yet another mechanism by which ETV6 can be implicated in the pathogenesis of hematologic malignancies. For a complete description of all aberrations involving ETV6 we refer the reader to a detailed review written by Stefan Bohlander.17

Translocations involving PDGFRB

When the ETV6-PDGFRB fusion gene was identified in 1994, it was not immediately clear that this would have tremendous therapeutic implications. However, during the development of imatinib, it was observed that this small molecule inhibitor of BCR-ABL also inhibited the activity of PDGFRα and PDGFRβ. Perhaps the most important consequence of this observation was the clinical application of imatinib for the treatment of chronic eosinophilic leukemia with the FIP1L1-PDGFRA fusion, and CMML/myeloproliferative diseases with ETV6-PDGFRB or variant PDGFRB fusions.4,18

To date, 11 additional fusion partners of PDGFRB have been identified (Table 2), and most patients with ETV6-PDGFRB or variant PDGFRB fusion genes were reported to respond extremely well to imatinib therapy. In this issue of the journal, Walz and co-workers report the identification of three more novel PDGFRB fusion genes.3 Two of the three patients in their study had marked eosinophilia, as is frequently observed in patients with myeloproliferative diseases and PDGFRα or PDGFRβ fusions. All three patients also showed good responses to imatinib therapy, and achieved complete hematologic and molecular remissions.3

Table 2.Translocations involving the PDGFRB gene in myeloproliferative disorders.

In most of the chromosomal aberrations affecting PDGFRB, the breakpoint in the PDGFRB gene is located in the intron upstream of the exon encoding the transmembrane region of PDGFRβ. In these fusions, the fusion partner of PDGFRB is believed to provide a homodimerization domain, which is required to dimerize the kinase domain of PDGFRβ, subsequently leading to its catalytic activation. Although this has not always been experimentally verified, most of the fusion partners of PDGFRβ do indeed contain domains that are known to possess potential to form homodimers. Interestingly, the breakpoint in PDGFRB in the PRKG2-PDGFRB fusion reported in this issue of Haematologica/The Hematology Journal has an unusual breakpoint that falls within the exon of PDGFRB that encodes the juxtamembrane region (WW-motif).3 This breakpoint resembles the breakpoints in PDGFRA in chronic eosinophilic leukemia cases with the FIP1L1-PDGFRA fusion.18 In the case of the FIP1L1-PDGFRA fusion, the breakpoints in PDGFRA disrupt the exon encoding the juxtamembrane region, and as a consequence, the juxtamembrane region of PDGFRα is only partially present in the FIP1L1-PDGFRα fusion protein. We have recently shown that the interruption of the juxtamembrane region is strictly required to activate the kinase activity of FIP1L1-PDGFRα, whereas the presence of the full juxtamembrane region inhibits kinase activity and transforming potential of FIP1L1-PDGFRα. The predicted PRKG2-PDGFRβ fusion protein described here by Walz et al.3 also lacks part of the juxtamembrane region of PDGFRβ, which could indicate that in this particular fusion, the interruption of the juxtamembrane region is required for kinase activity of PRKG2-PDGFRβ.

Future perspectives

Whatever the mechanism of activation, the type of fusion, or the fusion partner, the most important message from the work by Walz and co-workers is that patients with a myeloproliferative disease and a chromosomal translocation involving the PDGFRB gene respond well to imatinib therapy.3 As it is clear that there are a large number of translocations involving PDGFRB, with a different partner gene being involved each time, it may not be strictly required for diagnostic purpose to identify the fusion partner. It may be sufficient in these patients to prove by fluorescence in situ hybridization that PDGFRB is rearranged. It should be noted, however, that the identification of the exact fusion gene remains extremely valid for both research purposes and to enable molecular follow-up of the response of a patient to therapy. PDGFRB may have a license to fuse, imatinib still has its license to kill the leukemic cells with PDGFRB fusions. We will need to follow-up patients to determine whether imatinib as single agent therapy is sufficient to achieve long-term responses in patients with myeloproliferative diseases and translocations involving PDGFRB. If not, a number of other small molecule inhibitors that inhibit ETV6-PDGFRβ kinase activity at low nanomolar concentrations have already been identified.20,21