The proto-oncogene Fli-1 was first identified as a common proviral insertion site (Friend leukemia insertion site 1), observed in 75% of erythroleukemia cell clones induced by Friend murine leukemia virus (MuLV) in mice (3
). It encodes a transcription factor belonging to the ETS family, the founding member of which is the v-ets
oncogene, an oncogenic version of c-ets1
, transduced by the E26 leukemogenic virus (26
). More than 30 different members of the ETS family have been identified to date (25
). All members of this family share a highly conserved ETS DNA binding domain responsible for the fixation to a core consensus purine-rich sequence, GGA(A/T), found in a wide variety of viral and cellular transcriptional regulatory regions (28
). ETS proteins are involved in the regulation of many different biological processes ranging from morphogenesis and eye development in Drosophila melanogaster
to hematopoietic differentiation in mammals. In addition, many genes of the ets family participate in various oncogenic processes when activated as a result of chromosomal translocations or proviral insertions (see reference 13
for a review).
In addition to Friend erythroleukemia, activation of the Fli-1 gene by proviral insertion has been reported in several non-B and non-T leukemias induced by the Cas-Br-E virus (5
) and in granulocytic leukemia induced by the Graffy MuLV (9
). In all these cases, the proviral insertions occur in the 5′ region of the Fli-1 gene and are responsible for the overexpression of a normal FLI-1 protein. The Fli-1 gene also is rearranged in a majority of cases of the Ewing family of tumors that share t(11;22) chromosome translocation (8
). In these cases, the translocation is responsible for the production of an abnormal fusion protein, EWS–FLI-1, which harbors the N-terminal part of the EWS protein and the C-terminal part of FLI-1, including the ETS DNA binding domain, and which displays altered transactivation properties compared to normal FLI-1 (1
). Recently, we showed that the transcription of the Fli-1 gene is positively regulated by the SPI-1/PU.1 transcription factor (43
), another ETS protein involved in erythroleukemia induced by the spleen focus-forming virus component of the Friend viral complex (35
). We also showed that overexpression of FLI-1 inhibits the chemically induced erythroid terminal differentiation of spleen focus-forming virus-infected cells (43
). Similarly, it has been shown that overexpression of FLI-1 also inhibits the erythropoietin-dependent erythroid differentiation of one other mouse erythroleukemia cell line (44
) as well as avian primary erythroblasts (39
). In this latter case, inhibition of avian erythroid differentiation by FLI-1 is associated with increased proliferation, reduced apoptosis upon erythropoietin withdrawal, and deregulation of cyclin D2 and D3 gene expression (39
). The FLI-1 protein also displays antiapoptotic activity in NIH 3T3 cells (50
) and functionally interferes with nuclear hormone receptors (7
). The molecular targets of FLI-1 involved in all these processes remain unknown.
The Fli-1 gene is normally expressed in cells of various types during the early development of Xenopus
), mice (33
), and chickens (29
). In all three species, the Fli-1 gene is expressed in endothelial and neural crest cells. Further detailed studies with chickens (29
) have shown that Fli-1 gene expression in neural crest cells is restricted to mesenchymal lineages derived from neural crest cells at the end of their migration. The Fli-1 gene is also expressed in cartilage cells derived from mesoderm and could be expressed in the putative precursor of hematopoietic cells and angioblasts (29
transgenic mice have been established which, surprisingly, display a very discrete phenotype characterized by thymic hypocellularity (33
). However, these Fli-1−/−
mice are still able to produce an abnormal FLI-1TP
protein, modified in its N-terminal part, which could substitute for the normal FLI-1 protein functions. Overexpression of FLI-1, obtained by injection of synthetic Fli-1 mRNA into Xenopus
embryos, leads to dramatic development anomalies of the anteroposterior and dorsoventral polarities, of tissue differentiation, particularly in eye and cartilage development, and of erythroid differentiation including ectopic erythroid differentiation and hemangioma (42
). In contrast, very subtle anomalies limited to the abnormal proliferation of B-lymphoid cells have been observed in transgenic mice displaying a twofold overexpression of FLI-1 (52
). The unexpectedly discrete phenotype of these transgenic mice is most probably due to a FLI-1 overexpression lower than that obtained in Xenopus
embryos as well as to counterselection of transgenic embryos expressing higher levels. Taken together, these data strongly suggest that the Fli-1 gene most probably plays important roles in multiple differentiation and cell migration processes during development, which remain to be precisely identified.
Despite multiple unresolved questions concerning the normal functions of the Fli-1 gene and its role in leukemia, it is already clear that its expression needs to be tightly controlled in vivo. To date, three different promoters of the Fli-1 gene have been identified. One of these promoter was first identified in erythroleukemia cell lines induced by Friend MuLV. It is responsible for transcription initiating at position −398 upstream from the ATG triplet specifying the translation initiation of Fli-1 mRNA (2
). A second promoter has been identified 1.4 kb upstream. This promoter directs the synthesis of an alternatively spliced Fli-1 mRNA, named Fli-1b, which lacks exon 1 due to direct splicing between new exon 1b and exon 2 (10
). This alternative Fli-1b mRNA, which to date has been detected only in two human leukemic pre-B-cell lines, is responsible for the synthesis of a short FLI-1 protein by translation initiating in exon 2 instead of exon 1. More recently, we have identified a third promoter which is responsible for transcription initiating at position −204 and whose activity is positively regulated by transcription factor SPI-1/PU.1 (43
). Given the specific expression of Spi-1 in the hematopoietic tissue, this latter promoter might be responsible for the specific expression and function of the Fli-1 gene in hematopoiesis, but this possibility remains to be demonstrated.
All known Fli-1 mRNA isoforms harbor more than 200 nucleotides (nt) of 5′ untranslated region (5′ UTR). This suggests strongly that the 5′ UTR is involved in additional controls of Fli-1 gene expression at the posttranscriptional or translational level (19
). The present study was undertaken to investigate this latter possibility.