A Fli in the ointment

Comment on Masuya et al, page 95

Transcription factors that are aberrantly activated in leukemias often play a seminal role in blood lineages other than the malignant cell population. Knowing the full spectrum of transcription factor actions should inform our understanding of leukemia pathogenesis.

Leukemogenic chromosomal translocations commonly perturb transcription factor (TF) genes that are normally expressed in a restricted number of blood cell lineages. Many studies in the 1990s demonstrated the physiologic roles of these TFs in defining the character of their host cell types.¹  Subsequent investigation revealed additional roles in other lineages and surprising interactions among factors with ostensibly distinct functions.²  Indeed, TFs historically associated with a single or few closely related cell types may also be expressed, albeit at different levels or to different ends, in other lineages and especially in pluripotent progenitors. It is still unclear which class of TF activities is responsible for the leukemias that result from their misexpression.

In this issue of Blood, Masuya and colleagues add a new twist to an old story on experimental erythroleukemia in mice. Most malignant clones induced by integration of the Friend leukemia virus activate expression of Friend leukemia integration 1 (Fli-1), an Ets-family TF that may regulate transcription of many genes in the sibling erythroid and megakaryocytic lineages.³  Mice lacking Fli-1 die before definitive hematopoiesis occurs in development, though it is possible to detect defects in megakaryocyte and erythoid maturation. Nuanced changes in other blood cells, by contrast, were less evident and it is this aspect that Masuya and colleagues address in chimeric mice they created using both embryo aggregation and conventional bone marrow transplantation. In their experiments, cells derived from Fli-1^–/– or Fli-1^+/– embryos made a measurably smaller contribution toward circulating neutrophils and monocytes and possibly generated more natural killer (NK) cells than expected. The results suggest unanticipated functions for Fli-1 in aspects of hematopoiesis besides megakaryocyte differentiation; the experimental design presumes that defects arising in the absence of Fli-1 are intrinsic to the affected lineage.

Studies of this kind constitute an essential step toward full understanding of TF functions. The conclusions of the Masuya et al study, however, rest on analysis of very few animals and the effect of Fli-1 deficiency on leukocytes seems subtle and confined to the peripheral blood population: cell proportions and morphology in the bone marrow are unaffected. Moreover, the authors postulate a dependence on Fli-1 gene dosage that is not revealed independently in germline heterozygote mice. Each of these caveats will need to be addressed in future studies. Meanwhile, the results invite immediate consideration of the implications for erythroleukemia pathogenesis.

Different Friend virus strains induce similar hematologic changes and integrate near remarkably few gene loci. Besides Fli-1, another Ets-family TF, Spi-1 or Pu.1, is commonly involved, as is the p45 subunit of nuclear factor–erythroid 2 (NF-E2). Can common themes among these TFs shed light on the basis for erythroid progenitor hyperproliferation? Certainly, megakaryocytes manifest requirements for both Fli-1 and NF-E2 to complete their differentiation, though intuition might suggest those specific roles to be irrelevant for leukemogenesis, per se. Whereas PU.1 once seemed to be required principally for lymphomyeloid differentiation, recent studies raise the possibility of additional activities that distinguish myeloid from erythroid progenitors.^2,4  Confusion in the arena of lineage choice may underlie many neoplastic disorders, and the new findings about the versatile role of Fli-1 in granulocyte differentiation highlight this notion further.