The AML1 gene, located on chromosome 21, is involved in several distinct chromosomal translocations in human leukemia. In t(8;21) acute myelogenous leukemia (AML), the AML1 gene is juxtaposed to the ETO gene located on chromosome 8, generating an AML1͞ETO fusion protein. Both AML1͞ETO and the AML1 proteins recognize the same consensus DNA-binding motif (TGT͞CGGT), which is found in the promoters of several genes involved in hematopoiesis. We found that two myeloid leukemia cell lines with the t(8;21) translocation, Kasumi and SKNO-1, have elevated levels of BCL-2 protein compared with other myeloid cell lines. In addition, we identified a consensus AML1 binding site in the BCL-2 promoter. Thus far, AML1͞ETO has been shown to dominantly repress its target genes; however, we found that AML1͞ETO activates transcription of the BCL-2 gene in U937 cells. This activation requires the presence of both the runt homology domain (rhd) and the C-terminal portion of AML1͞ ETO. We demonstrated sequence specific binding of both AML1A and AML1͞ETO to the TGTGGT sequence in the BCL-2 promoter and showed that the AML1 binding site is required for responsiveness to AML1͞ETO. Interestingly, AML1A and AML1B do not modulate the activity of the BCL-2 promoter. The elevated levels of BCL-2 in cells that express AML1͞ETO may prolong their life span and contribute to the development of t(8;21) leukemia.Chromosomal translocations found in human leukemia frequently involve genes that code for transcription factors (1). In acute myeloid leukemia (AML) with the t(8;21) chromosomal translocation, which occurs in Ϸ40% cases of AML with the M2 French-American-British subtype, coding sequences of the AML1 gene (on chromosome 21) are juxtaposed to coding sequences of the ETO gene (on chromosome 8) generating an AML1͞ETO fusion protein (2, 3). The AML1 family of transcription factors recognize the binding sequence 5Ј-TGT͞ CGGT-3Ј (2) through an 117-amino acid region that is highly homologous to the Drosophila segmentation gene runt (2, 3), and has been called the runt homology domain (rhd). This domain is necessary for DNA binding, as well as for proteinprotein interactions (3). At least three forms of AML1 protein are produced by alternative splicing (4). The AML1-B isoform (479 amino acids) contains the rhd and a putative C-terminal transcriptional activation domain; the AML1-A isoform (250 amino acids) contains the DNA binding domain, but lacks the potential transcriptional activation domain. AML1-B, but not A ML-1A, can transactivate the human granulocyte͞ macrophage colony-stimulating factor (GM-CSF) promoter (5) and the T cell receptor ␤ enhancer (6), whereas both isoforms can transactivate the human interleukin 3 (IL-3) gene (H.U., S.Z., and S.D.N., unpublished work).The genes encoding AML1 or its dimerization partner CBF␤, have been shown to be involved in several other translocations in human acute leukemia (7). The AML1 gene is fused to the TEL gene in t(12;21) acute lymphoblastic leukemia (8). In the t(3;21) translocation, seen ...

No abstract

T-cell activation induces expression of the hematopoietic growth factor granulocyte-macrophage colony-stimulating factor (GM-CSF). To define the molecular events involved in the induction of GM-CSF gene expression more clearly, we prepared and analyzed deletion mutants of GM-CSF promoter recombinant constructs. The results localized inducible expression to a 90-base-pair region (-53 to +37) which is active in uninfected and human T-cell leukemia virus-infected T-cell lines but not in resting or mitogen-stimulated B cells. DNase I footprinting experiments revealed protection of sequences contained within this region, including a repeated nucleotide sequence, CATT(A/T), which could serve as a core recognition sequence for a cellular transcription factor. Upstream of these GM-CSF promoter sequences is a 15-base-pair region (-193 to -179) which has negative regulatory activity in human T-cell leukemia virus-infected T cells. These studies revealed a complex pattern of regulation of GM-CSF expression in T cells; positive and negative regulatory sequences may play critical roles in controlling the expression of this potent granulopoietin in the bone marrow microenvironment and in localized inflammatory responses.

The hematopoietic growth factor GM-CSF (granulocyte-macrophage colony-stimulating factor) is expressed by activated but not resting T lymphocytes. Previously, we localized GM-CSF-inducible promoter activity to a 90-bp region of GM-CSF 5'-flanking sequences extending from bp -53 to +37. To more precisely identify the GM-CSF DNA sequences required for inducible promoter activity in T lymphocytes, we have performed mutagenesis within a region of GM-CSF 5'-flanking sequences (bp -57 to -24) that contains the repeated sequence CATT(A/T). Mutations that do not alter the repeated CATT(A/T) sequence do not eliminate inducible promoter activity, whereas mutation or deletion of either of the CATT(A/T) repeats eliminates all inducible promoter activity in T-cell lines and in primary human T lymphocytes. Thus, both copies of the direct repeat CATT(A/T) are required for mitogen-inducible expression of GM-CSF in T cells.

The frequent occurrence of TF gene involvement in translocations associated with leukemia is remarkable, although not yet explained. The wide variety of TFs involved in these translocations and the different stages of cellular maturation argue against a unifying mechanism. Recombinases, active during B-cell and T-cell development, have been implicated in gene arrangements involving TCR genes and in the SIL/SCL rearrangement, which involves two genes not normally rearranged. However, other mechanisms must clearly be active in generating these molecular abnormalities and perhaps they relate to the multistep maturation and differentiation processes and continuous cell turnover seen in hematopoietic cells. The difficulties in obtaining human solid tumor samples may make it more difficult to identify translocations involving TF genes in solid tumors. Recently, the cytogenetic analysis of solid tumors has improved and specific cytogenetic abnormalities have been associated with specific types of tumors. With advanced techniques, such as fluorescent in situ hybridization (a technique that does not depend on cell growth) and PCR, abnormalities involving TF genes will be discovered. Abnormalities of TF genes, other than translocations, have been seen in a broad variety of nonhematopoietic malignancies. The p53 protein has been shown to bind DNA in a sequence- specific fashion and interact with a variety of DNA tumor virus oncoproteins. The broad range of cell types that harbor p53 abnormalities suggests that TF abnormalities will likely be implicated in many solid tumors. We have detailed several examples of how gene rearrangements that accompany chromosomal translocations in acute leukemia can alter the expression or activity of cellular TFs. Several translocations generate fusion RNA transcripts and fusion TF proteins with altered functional characteristics. Other translocations result in the expression of a gene not normally detectable in hematopoietic cells or alter the level of its expression, or affect the promoter usage or exon structure of the gene (Table 2). Studies are underway in many laboratories to characterize the biologic activity of these abnormal TFs and it remains to be proven that these molecular abnormalities are directly linked with leukemogenesis. The identification of abnormal fusion transcripts and proteins may allow specific therapies to be directed against “tumor-specific” DNA, mRNA, or protein targets. Therapeutic strategies based on antisense or ribozyme technology may be used to turn off expression of these genes and inhibit leukemia cell growth. Immunologic methods can also be used to direct therapy against the malignant cells.