The v-Myb oncogene causes monoblastic leukemia and transforms only myelomonocytic cells in culture. The v-Myb protein is nuclear and binds to specific DNA sequences. To identify genes regulated by v-Myb, we utilized primary cells transformed by a retrovirus encoding a v-Myb-estrogen receptor (ER) fusion protein. The Ets-2 gene was not expressed in v-Myb-ER transformed cells in the presence of estradiol, but was expressed within 4 h after estradiol withdrawal. The expression of Ets-2 also increased dramatically following phorbol ester-induced differentiation of the v-Myb-transformed BM2 cell line. Conversely, CRYP-alpha, encoding a transmembrane tyrosine phosphatase, was expressed in the presence but not the absence of estradiol in v-Myb-ER transformed cells. CRYP-alpha was downregulated during the phorbol ester-induced differentiation of BM2 cells. Although LIM-3 expression was estradiol-inducible in v-Myb-ER transformed monoblasts, LIM-3 was expressed neither in primary yolk sac cells transformed by unfused v-Myb nor in BM2 cells. We conclude that although v-Myb has been intensively studied as a transcriptional activator, v-Myb can repress biologically relevant genes such as Ets-2, which promotes macrophage differentiation. In addition, we have shown that some genes that are regulated by a v-Myb-ER fusion protein may not be relevant to the biological function of the unfused v-Myb protein.
The v-Myb oncogene of the avian myeloblastosis virus is unusual in that it causes only monoblastic leukemia in chickens and transforms only myelomonocytic cells in culture (Lipsick and Wang, 1999). Studies from a number of different laboratories have shown that the v-Myb protein is nuclear, binds to specific DNA sequences as a monomer, and can activate the expression of reporter genes (Boyle et al., 1984; Klempnauer et al., 1984; Biedenkapp et al., 1988; Ibanez et al., 1988; Garcia et al., 1991). In addition, both v-Myb and its normal cellular homologue c-Myb have been shown to be capable of activating gene expression in a variety of cell types (Ness et al., 1989; Nishina et al., 1989; Weston and Bishop, 1989; Ibanez and Lipsick, 1990; Nakagoshi et al., 1992; Siu et al., 1992; Cogswell et al., 1993; Ku et al., 1993; Melotti and Calabretta, 1994; Burk et al., 1997; Worpenberg et al., 1997; Schlichter et al., 2001; Chen et al., 2002; Bartley et al., 2003; Rushton et al., 2003; Braas et al., 2004; Lang et al., 2005; Liu et al., 2006).
To attempt to identify biologically relevant Myb-regulated genes, we previously described a system in which primary chicken yolk sac cells could be oncogenically transformed in an estrogen-dependent fashion by a retrovirus that encoded a fusion of AMV v-Myb and the hormone-binding domain of the human estrogen receptor (Engelke et al., 1997). We found that in the absence of estrogen, these transformed cells predominantly differentiated into multinucleated giant cells. By modifying the culture conditions with the addition of recombinant chicken myeloid growth factor (Leutz et al., 1984), a G-CSF-related cytokine, we were able to grow relatively large numbers of primary v-Myb-ER transformed monoblasts that could be differentiated into mononuclear macrophages by the simple withdrawal of estradiol (Figure 1). The presence of macrophage differentiation was readily apparent within 24 h of hormone withdrawal and by 48 h virtually all cells in the culture showed features of macrophage differentiation including increased adherence to substrate, the induction of phagocytic vacuoles, and cessation of proliferation.
In order to identify genes that might be directly regulated by v-Myb, we extracted mRNA from v-Myb-ER transformed primary monoblasts maintained in estradiol and from sister cultures four hours after the withdrawal of estradiol. To determine which technique might be most useful for differential analysis of these mRNA populations, we conducted pilot studies on the v-Myb transformed BM2 cells line which can be induced to differentiate into mature macrophage following treatment with phorbol ester (Pessano et al., 1979; Moscovici et al., 1982). We found that in our hands, representation display analysis (RDA) (Hubank and Schatz, 1994) was quite reproducible and readily identified both known and unknown phorbol-inducible genes including collagenase, interleukin 8, ATP synthase, and macrophage inflammatory protein, all of which were verified by Northern blotting (Supplemental Data, Figure S1). We therefore utilized the RDA method with three rounds of differential amplification of cDNA from v-Myb-ER transformed primary monoblasts before versus 4 h after withdrawal of estradiol (Figure 2). Subtraction was performed in both directions in order to identify genes whose expression was enriched following the withdrawal of estradiol, as well as those genes whose expression was enriched in the presence of estradiol.
Each visible ethidium bromide stained band from the third round of differential amplification was excised, extracted, molecularly cloned, sequenced and analyzed by a BLAST search of the NCBI databases. A number of genes were found to be significantly enriched in each pool (Table 1). We then used RNase protection assays as a further screen to identify those genes that were differentially regulated by four hours of estradiol withdrawal in v-Myb-ER transformed primary cells. We chose to focus on two genes that were repressed by v-Myb-ER in that their expression increased dramatically after withdrawal of estradiol (ets-2 and chemokine ah221), and on two genes that were induced by v-Myb-ER (CYRP-alpha tyrosine phosphatase and lim-3) (Figure 3). In order to better understand the biological relevance of these genes to the transformed phenotype, we performed RNase protection assays in two additional systems. First, we compared gene expression in primary cells transformed by v-Myb without an estrogen receptor fusion. As predicted, these cells do not differentiate in response to estradiol. Second, we compared gene expression in the v-Myb transformed BM2 cell line before and after four hours of treatment with the differentiation-inducing agent phorbol ester.
Three classes of genes were identified by these experiments. Perhaps most interesting is our observation that ets-2, which itself encodes a DNA-binding transcription factor, appears to be repressed by v-Myb. In particular, ets-2 expression was absent in v-Myb-ER transformed primary cells but was induced within four hours after withdrawal of estradiol. Consistent with these observations, ets-2 expression was absent in v-Myb transformed primary cells in the presence or absence of estradiol. In addition, the levels of ets-2 increased dramatically in BM2 cells upon treatment with phorbol ester, an induced of macrophage differentiation. Previous work has shown that Myb proteins are capable of repressing the transcription of transfected reporter genes. Our results now identify ets-2 as a physiologically relevant target for repression by v-Myb. Our screen also identified chemokine ah221 as a gene repressed by v-Myb-ER and v-Myb in primary leukemic cells and in the BM2 cell line (Figure 3).
Several lines of additional evidence imply that ets-2 is directly regulated by v-Myb and that repression of ets-2 expression contributes to the highly virulent leukemogenic phenotype that was selected during the passage of AMV from chicken to chicken for many years. First, the myeloid-specific regulatory sequences of ets-2 contain Myb-binding sites that have been conserved during the evolution of human and chicken (Begue et al., 1997). Second, chromatin immunoprecipitation experiments using anti-Myb specific antibodies demonstrated that the v-Myb protein was directly bound to the ets-2 promoter in v-Myb-transformed BM2 cells (Figure 4). Third, the ectopic expression of ets-2 induces differentiation in a murine myeloid leukemia cell line (Aperlo et al., 1996). Fourth, a hypomorphic allele of ets-2 greatly reduced myeloid proliferation and inflammation in the motheaten-viable (me-v) mouse model that is caused by germline mutation of the hematopoietic cell tyrosine phosphatase (Hcph/SHP1) (Wei et al., 2004).
The observation that ets-2 is expressed to some degree in undifferentiated BM2 cells is consistent with previous work showing that BM2 cells are not a reliable model for primary v-Myb-transformed monoblasts. For example, expression of the mim-1 gene has never been detected in primary cells transformed by AMV v-Myb, but is readily detected in undifferentiated BM2 cells (Dini et al., 1995; Ness et al., 1989). In this regard, we created a variant of the BM2 cell line containing a tetracycline-regulated ets-2 cDNA but were unable detect any significant changes in proliferation, differentiation, or survival following induction of additional ets-2 expression (data not shown).
Interestingly, the c-myb proto-oncogene has been independently transduced in two different avian acute leukemia viruses (Lipsick and Wang, 1999). AMV encodes v-Myb, a doubly truncated form of the normal c-Myb protein that also contains 10 amino-acid substitutions, which also contribute to leukemogenesis. In contrast, the E26 leukemia virus encodes a fusion protein in which a smaller portion of the c-Myb protein with a single amino acid substitution is fused directly to a large segment of the c-Ets-1 transcription factor. Previous studies have shown that the portion of c-Myb present in E26 is at best weakly oncogenic (Metz and Graf, 1991). Furthermore, the passage in vivo of a non-leukemogenic virus that encodes separate E26 Myb and c-Ets-1 proteins results in the selection for leukemogenic viruses in which deletions recreate a Myb-Ets fusion protein. The results presented above suggest that this fusion may be under selection for its ability to repress targets of both Myb and Ets transcription factors, perhaps because E26 Myb itself is not an efficient repressor of ets-2 transcription.
A second class of gene identified in our screen is represented by CYRP-alpha, which encodes a transmembrane tyrosine phosphatase. This gene is expressed in primary monoblasts transformed by v-Myb-ER in the presence of estradiol, but is dramatically downregulated within four hours of withdrawal of estradiol (Figure 3). CYRP-alpha is also expressed in primary monoblasts transformed by unfused v-Myb, in either the presence or absence of estrdiol. These results demonstrate that v-Myb-ER rather than endogenous ER is responsible for the activation of CYRP-alpha in v-Myb-ER transformed monoblasts. In addition, CYRP-alpha was highly expressed in the BM2 cell line. However, phorbol ester treatment did not inhibit CYRP-alpha expression in BM2 cells, implying that the down-regulation of this gene is not required for macrophage differentiation. We note that unlike primary leukemic cells, phorbol ester treatment of BM2 cells is readily reversible. These results therefore suggest that genes critical for cell proliferation may be deregulated in this cell line by additional mutations and/or epigenetic alterations.
A third class of gene identified in our screen is represented by lim-3, which encodes a homeobox-containing transcription factor critical for neural development. lim-3 is expressed in v-Myb-ER transformed primary monoblasts in the presence of estradiol, but is rapidly down-regulated after withdrawal of estradiol in a fashion similar to CYRP-alpha. Surprisingly, no lim-3 expression was detected in primary monoblasts transformed by v-Myb itself or in the BM2 cell line. These results imply that activation of lim-3 expression by the v-Myb-ER fusion protein represents a neomorphic phenotype caused by fusion of the estrogen receptor hormone-binding domain.
In order to identify genes directly regulated by v-Myb and c-Myb, a variety of experimental approaches have been used. The first Myb-regulated gene, mim-1, was identified in a differential cDNA library screen of primary cells transformed by a temperature sensitive mutant of the E26 virus, which encodes a tripartite Gag-Myb-Ets fusion protein (Ness et al., 1989). However, although mim-1 is abundantly expressed in differentiated myeloid cells, this gene was not expressed in monoblasts transformed by AMV v-Myb, nor were other Myb-regulated genes identified in this screen.
In an approach similar to that which we have taken here, a hybrid AMV/E26 v-Myb protein was fused to the hormone-binding domain of the human estrogen receptor and introduced into an established v-Myc-transformed macrophage cell line (Burk and Klempnauer, 1991). The utility of this system has been proven by the identification of a number of Myb-regulated genes (Burk et al., 1997; Worpenberg et al., 1997; Schlichter et al., 2001; Braas et al., 2004). A drawback of this system is that the cells of interest are already transformed by the v-Myc oncogene and are therefore not dependent upon the v-Myb-ER fusion protein for growth or survival. Myb-regulated genes have also been identified in murine cell lines transformed by a truncated c-Myb-ER fusion protein (Ferrao et al., 1997; Bartley et al., 2003). An advantage of this system is that these cells are dependent both upon Myb and upon exogenous growth factors for their proliferation and survival.
An alternative approach has been to create dominant interfering forms of Myb proteins by the fusion of the c-Myb DNA-binding domain, the transcriptional repressor domain from the Drosophila Engrailed protein, and the human estrogen receptor which were then expressed in established hematopoietic cell lines (Taylor et al., 1996; White and Weston, 2000). This system has recently been used in conjuction with a subtractive cDNA cloning method to identify a number of Myb-regulated genes (Lang et al., 2005). One disadvantage of this system is that one cannot distinguish among inhibition of different Myb proteins (e.g. c-Myb versus B-Myb). A similar method has been described in which a Myb-Engrailed fusion protein has been inducibly expressed in a transformed erythroleukemia cell line under control of the metallothionein promoter (Chen et al., 2002).
Recently, another approach has been described in which established cell lines of various lineages are infected with adenovirus vectors that express a variety of different wild-type and mutant Myb proteins and then gene expression has been analysed by microarray (Rushton et al., 2003; Lei et al., 2004; Liu et al., 2006). The most remarkable finding of these studies has been that the panel of Myb-regulated identified in such experiments is highly dependent upon the specific Myb protein expressed as well as the specific cell line used for adenoviral infection. In this regard, those genes we have identified as being directly regulated by a v-Myb-ER fusion protein in primary transformed chicken monoblasts do not appear to have been identified in several other screens in which Myb proteins were expressed in various other cell types (Rushton et al., 2003; Lang et al., 2005; Liu et al., 2006).
In summary, we have shown that in addition to its well-characterized role as an activator of gene expression, the v-Myb oncoprotein represses physiologically important genes including ets-2. We have also shown that v-Myb-ER fusion proteins can regulate genes not regulated by v-Myb itself. Therefore, caution must be used in interpreting the results of experiments in which fusions of the estrogen receptor hormone-binding domain and other regulatory domains have been used to identify genes regulated by Myb and other DNA-binding proteins.
This work was supported by USPHS research grant R01 CA43592 (JSL) and by training grant T32 CA09151 (DMW).