Cyclin E is a target of WT1 transcriptional repression.

WT1 was originally identified as a Wilms' tumor suppressor gene, but it may have oncogenic potential in leukemia and in some solid tumors. WT1 is a transcription factor that has been implicated in the regulation of target genes related to apoptosis, genitourinary differentiation, and cell cycle progression. Because induction of WT1 leads indirectly to increased p21 expression in osteosarcoma cells, we investigated the possibility that other genes involved in the G(1)/S phase transition might also be WT1 targets. Cyclin E plays a crucial role in the cell cycle by activating cyclin-dependent kinase 2, which phosphorylates Rb, leading to progression from G(1) into S phase. We identified several WT1 binding sites in the cyclin E promoter. We demonstrate that WT1 binds to these sites and that in transient transfection assays WT1 represses the cyclin E promoter. This activity is dependent on the presence of a binding site located downstream of the transcription start site. In intact cells, induction of WT1 expression down-regulates cyclin E protein levels. These results provide the first demonstration that WT1 can directly modulate the expression of a gene involved in cell cycle progression.

From The Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University, Baltimore, Maryland 21231 WT1 was originally identified as a Wilms' tumor suppressor gene, but it may have oncogenic potential in leukemia and in some solid tumors. WT1 is a transcription factor that has been implicated in the regulation of target genes related to apoptosis, genitourinary differentiation, and cell cycle progression. Because induction of WT1 leads indirectly to increased p21 expression in osteosarcoma cells, we investigated the possibility that other genes involved in the G 1 /S phase transition might also be WT1 targets. Cyclin E plays a crucial role in the cell cycle by activating cyclin-dependent kinase 2, which phosphorylates Rb, leading to progression from G 1 into S phase. We identified several WT1 binding sites in the cyclin E promoter. We demonstrate that WT1 binds to these sites and that in transient transfection assays WT1 represses the cyclin E promoter. This activity is dependent on the presence of a binding site located downstream of the transcription start site. In intact cells, induction of WT1 expression down-regulates cyclin E protein levels. These results provide the first demonstration that WT1 can directly modulate the expression of a gene involved in cell cycle progression.
WT1 was originally identified as a tumor suppressor gene in hereditary cases of Wilms' tumor. This gene encodes a 57-kDa protein with an amino-terminal transcriptional regulatory domain and a carboxyl-terminal zinc finger DNA binding domain (1,2). WT1 mRNA is subject to two alternative splicing events leading to the generation of four distinct transcripts (3). The first alternative splice involves exon 5, which is either included in or excluded from the mature message. The other alternative splice involves a choice between two 3Ј splice acceptor sites at the beginning of exon 10. Selection of the more 5Ј splice acceptor site adds nine base pairs (referred to as the KTS insert for the 3 amino acids encoded by these base pairs) to exon 10. These 3 additional amino acids alter the spacing between the third and fourth zinc fingers, changing the DNA recognition site of the protein (4,5). The subcellular localization of WT1 and its association with RNA splicing factors are also affected by the presence or absence of the KTS insert (6,7).
There are two classes of genes regulated by WT1. The first of these is composed of genes critical for the differentiation of the specific cell types that express WT1, in particular, the regulation of sex determination. For example, the Dax-1 promoter is dramatically up-regulated by the WT1 isoform lacking both exon 5 and the KTS insert (designated WT1(Ϫ/Ϫ)) and by the isoform that contains exon 5 but lacks the KTS insert (designated WT1(ϩ/Ϫ)), whereas the isoforms containing the KTS insert (designated WT1(Ϫ/ϩ)) and both exon 5 and the KTS insert (designated WT1(ϩ/ϩ)) have little effect (8). A similar pattern is seen in the regulation of the mullerian inhibitory substance and SRY promoters by WT1 (9, 10). Each of these target genes is important in differentiation of the genitourinary system.
The second class of genes regulated by WT1 includes those involved in cell cycle regulation and apoptosis. The cyclin-dependent kinase inhibitor p21 is up-regulated when WT1(Ϫ/Ϫ) is induced, although there is no evidence that this is a direct effect of WT1 on the p21 promoter (11). Bcl-2 is also up-regulated by WT1(Ϫ/Ϫ), and this is a direct result of WT1 binding to sites in the bcl-2 promoter (12). These genes, despite serving more general roles in cellular function than Dax-1 and SRY, might also be important in terminal differentiation. The identification of p21 as a WT1-responsive gene led us to hypothesize that WT1 might also regulate the expression of other mediators of cell cycle progression. Here we present evidence that WT1 inhibits the expression of cyclin E and that the cyclin E promoter is a direct target of transcriptional regulation by WT1.
Plasmids-pCB6WT1(Ϫ/Ϫ) contains the WT1(Ϫ/Ϫ) cDNA cloned downstream of the CMV immediate-early promoter, and pCB6WT1(ϩ/ϩ) contains the cDNA encoding WT1(ϩ/ϩ). Plasmid 2-3 (a kind gift from Dr. R. Weinberg, Whitehead Institute, Cambridge, MA) contains 2.2 kb of the human cyclin E promoter and first exon cloned into the multiple cloning site of pGL2basic, driving the expression of the cDNA for firefly luciferase (Promega, Madison, WI). The plasmid pRL-SV40 contains the cDNA for Renilla luciferase under the control of the SV40 early enhancer/promoter region.
Transient Transfection Assays-Transient transfection assays were performed using TransIT LT-1 (PanVera, Madison, WI) following the manufacturer's protocol. NIH 3T3 cells were plated in 6-well plates at 10 5 cells/well. Each well was transfected with 2.5 g of plasmid 2-3, 0.7 g of either the WT1 construct or the pCB6 vector, and 0.1 g of pRL-SV40 as an internal control. 32D cl3 cells were plated in 100-mm tissue culture plates at 2 ϫ 10 5 cells/plate and were transfected with 10 * This work was supported by Grant R01-CA48943 from the National Institutes of Health (to S. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. g of plasmid 2-3. Luciferase assays were performed using the Dual Luciferase Reporter Assay System (Promega), and results were normalized by cotransfection with the pRL-SV40 Renilla luciferase plasmid.
Site-directed Mutagenesis-Mutagenesis of the WT1 cDNA was performed using the GeneEditor site-directed mutagenesis kit (Promega). The mutagenic oligonucleotide, 5Ј-GCGACGTGTTGCCTGGAGTAG-3Ј (inserted T is in bold type), was synthesized based on the published sequence of a WT1 mutation present in a patient (case 126) with acute myeloid leukemia (13). Successful mutagenesis was confirmed by DNA sequencing.
Electrophoretic Mobility Shift Assays-Nuclear extracts were prepared by a modification of the method of Schreiber et al. (14). Briefly, cells were washed with phosphate-buffered saline and resuspended in 10 mM HEPES, pH 7.9, 10 mM KCl, 0.2 mM EDTA, 1 mM dithiothreitol, 0.1 mg/ml phenylmethylsulfonyl fluoride, and 0.1 g/ml pepstatin A (all obtained from Sigma). The buffer contained additional protease inhibitors added by dissolving 1 complete Mini Protease Inhibitor Mixture tablet (Roche Molecular Biochemicals)/7 ml of buffer. Cells were incubated for 15 min at 4°C and then lysed by vortexing after the addition of Nonidet P-40 to a concentration of 0.5%. Nuclei were pelleted by centrifugation and resuspended in 20 mM HEPES, pH 7.9, 400 mM NaCl, 2 mM EDTA, 1 mM dithiothreitol, 0.1 mg/ml phenylmethylsulfonyl fluoride, and 0.1 g/ml pepstatin A with additional protease inhibitors as described above. Nuclei were extracted at 4°C for 15 min, and insoluble material was removed by centrifugation.
Promoter Mutagenesis-To generate expression constructs containing the cyclin E promoter lacking various WT1 binding sites, short deletions were made. To delete putative WT1 binding site 1 (nucleotides Ϫ745 to Ϫ737, the promoter was digested with BstXI (at position -868) and NdeI (at position -731), and the ends were made blunt with the Klenow fragment of DNA polymerase and religated together. To delete putative WT1 binding site 2 (nucleotides Ϫ507 to -499) the promoter was digested with BstEII (at position -559) and BlpI (at position -417), and the ends were made blunt with the Klenow fragment of DNA polymerase and religated. These constructs, missing both sites 1 and 2, were a gift from Dr. R. Weinberg (Whitehead Institute, Cambridge, MA) and contain a smaller fragment of the cyclin E promoter (positions -364 to ϩ 1007) controlling the expression of firefly luciferase. Deletion of putative WT1 binding site 3 (nucleotides ϩ262 to ϩ270) was accomplished by digesting the plasmids with AccI (at position ϩ204) and SrfI (at position ϩ368), making the ends blunt with the Klenow fragment of DNA polymerase, and religating. Promoter mutants containing deletions of multiple sites were generated by serial restriction digests and religations as described for the individual sites. All mutants were confirmed by DNA sequencing.

WT1(Ϫ/Ϫ) Regulates Cyclin E Expression-
Among the proposed targets of WT1 is the cyclin-dependent kinase inhibitor p21 (11). Because p21 participates in the G 1 /S transition, we hypothesized that other regulators of this phase of the cell cycle might also be affected by WT1. To examine this possibility, we utilized a stably transfected cell line, designated 32DWT1.1, expressing an inducible human WT1(Ϫ/Ϫ) cDNA. The details of this cell line are described elsewhere, 1 but briefly, the cell line was generated by transfecting 32D cl3 cells with a plasmid encoding the WT1(Ϫ/Ϫ) cDNA under the control of the sheep metallothionein promoter and selecting stable transformants in a mass culture. Treating 32DWT1.1 cells with 100 M ZnCl 2 overnight induces significant expression of the WT1(Ϫ/Ϫ) protein (Fig. 1). 32DWT1.1 cells and the control cell line 32DV4 (which was stably transfected with an empty metallothionein promoter-containing expression vector) were treated overnight with or without ZnCl 2 , and lysates were subjected to Western blotting with antibodies against a variety of proteins involved in cell cycle progression. In 32DWT1.1 cells, which express low levels of WT1(Ϫ/Ϫ) in the uninduced state, the basal level of p21 WAF1/CIP1 was less than the basal level in 32DV4 cells, but induction of WT1(Ϫ/Ϫ) did not alter p21 WAF1/CIP1 protein expression. This suggests that the ability of WT1 to induce p21 expression is cell type-specific. Induction of WT1(Ϫ/Ϫ) by ZnCl 2 also did not affect p27 Kip1 expression and had no effect on the expression of a number of other mediators of cell cycle progression, including Cdk2, Cdk3, Cdk4, cyclin D3, Cdc25a, JAK3 (Fig. 1), or cyclin D2 (data not shown). Under the same experimental conditions, however, the level of cyclin E protein in WT1(Ϫ/Ϫ)-expressing cells was reduced by almost 4-fold (Fig. 1). WT1 Suppresses the Cyclin E Promoter-Next, we sought to determine whether the decrease in the level of cyclin E protein upon induction of WT1(Ϫ/Ϫ) expression occurs through direct transcriptional suppression of the cyclin E promoter by WT1 (Ϫ/Ϫ). A reporter plasmid containing firefly luciferase under the control of the human cyclin E promoter was transfected into NIH 3T3 cells. Cotransfection of WT1(Ϫ/Ϫ) under the control of the cytomegalovirus immediate-early promoter resulted in a 4-fold decrease in luciferase activity (Fig. 2B). There was some variability in the degree of suppression in different experiments, with a range of 2.5-to 5-fold, although in most experiments there was a 4-to 5-fold suppression. In the same cells, the WT1(ϩ/ϩ) isoform was a much less efficient repressor of the cyclin E promoter, suppressing transcription by only 30% (Fig. 2B). A truncated form of WT1 lacking the carboxyl-terminal zinc finger region, which is required for DNA binding, had no effect on the cyclin E promoter in this assay (Fig. 2B). There was a clear dose-related suppression of the promoter by WT1(Ϫ/Ϫ), with decreasing luciferase activity as the ratio of WT1(Ϫ/Ϫ) to reporter plasmid increased (Fig. 2C). Because the ability of WT1 isoforms to modulate the activity of some promoters can vary by cell type, we wanted to determine whether the cyclin E promoter is also suppressed by WT1(Ϫ/Ϫ) in hematopoietic cells. Accordingly, we transiently transfected 32DWT1.1 cells with the cyclin E promoter-luciferase reporter construct and then treated the cells with ZnCl 2 to induce the expression of the transfected WT1(Ϫ/Ϫ). Just as in the NIH 3T3 cells, the WT1(Ϫ/Ϫ) suppressed the cyclin E promoter in the 32DWT1.1 cells (Fig. 2D). No suppression was seen in the control 32DV4 cells transfected with the same construct and treated with ZnCl 2 . These findings suggest that WT1(Ϫ/Ϫ) efficiently and specifically down-regulates cyclin E levels, possibly by directly binding to the promoter and repressing transcription.
Because WT1 has been shown to act as a dimer (15), we investigated the possibility that the WT1(ϩ/ϩ) isoform could interfere with the ability of the WT1(Ϫ/Ϫ) isoform to repress the cyclin E promoter in this assay. NIH 3T3 cells were therefore transfected with the cyclin E reporter plasmid and WT1 (Ϫ/Ϫ), WT1(ϩ/ϩ), or a combination of the two isoforms in two  ). B, C, and E, NIH 3T3 cells were cotransfected with this plasmid and either an empty expression vector (control) or the indicated WT1 isoform. Mutant refers to a deletion mutant that encodes a protein lacking the zinc finger region, which therefore does not bind DNA. D, 32DV4 and 32DWT1.1 cells were also transfected with the reporter construct and were treated with or without 100 M ZnCl 2 on the day of transfection. Relative luciferase was normalized to cotransfected Renilla luciferase activity using the Dual Luciferase Reporter Assay Kit. Error bars represent standard error of the mean as determined in triplicate assays. B, effect of different WT1 isoforms on the cyclin E promoter. C, dose-response curve showing the effect of increasing ratio of WT1(Ϫ/Ϫ) to cyclin E promoter. D, effect of inducible WT1(Ϫ/Ϫ) on cyclin E promoter activity in 32D cl3 cell lines. E, effect of co-transfecting WT1(Ϫ/Ϫ) and WT1(ϩ/ϩ). different ratios. WT1(Ϫ/Ϫ) again efficiently repressed the cyclin E promoter, whereas WT1(ϩ/ϩ) was much less effective. Remarkably, the degree of promoter repression was less with increasing ratio of WT1(ϩ/ϩ) to WT1(Ϫ/Ϫ) (Fig. 2E). This is unlikely to represent competitive DNA binding, because WT1(ϩ/ϩ) has only weak affinity for WT1(Ϫ/Ϫ) binding sites. More likely, this reflects the ability of WT1(ϩ/ϩ) to heterodimerize with WT1(Ϫ/Ϫ) and to thereby inhibit DNA binding by the WT1(Ϫ/Ϫ) isoform.
WT1(Ϫ/Ϫ) Binds to the Binding Sites in the Cyclin E Promoter-There are three potential binding sites in the cyclin E promoter for this isoform of WT1 ( Fig. 2A); two sites are 5Ј of the transcription start site, and one is within exon 1 (although 5Ј of the translation start site), an arrangement seen in other promoters, such as the insulin-like growth factor I receptor, that are repressed by WT1(Ϫ/Ϫ). To confirm that WT1(Ϫ/Ϫ) binds to the putative binding sites identified in the cyclin E promoter, we performed electrophoretic mobility shift assays. Nuclear extracts were made from 32DV4 cells and from stably transfected cells treated overnight with ZnCl 2 to induce a high level of WT1(Ϫ/Ϫ) expression. The presence of WT1 protein in the extracts from the transfected cells was confirmed by Western blotting (Fig. 3A). Nuclear extract from the control cells contains an activity that binds to and retards the migration of a radiolabeled oligonucleotide containing WT1 binding site 1 (Fig. 3B). Nuclear extracts from WT (Ϫ/Ϫ)-expressing cells also contain DNA binding activity that recognizes this oligonucleotide. In contrast to the control cells, nuclear extracts from the WT1-expressing cells generate two distinct species in this assay. One of these species comigrated with the complex derived from the 32DV4 cells, and the second complex migrated slightly more rapidly (Fig. 3B). Inclusion of an excess of unlabeled oligonucleotide in the reaction blocked the formation of both complexes. Strikingly, inclusion of an excess of an unlabeled oligonucleotide containing two point mutations that abrogate WT1 binding specifically eliminated the formation of the more slowly migrating complex, with no effect on the more rapidly migrating complex. These findings demonstrate that the slower migrating species represents a nonspecific protein-oligonucleotide complex, whereas the more rapidly migrating complex represents WT1(Ϫ/Ϫ) binding to the labeled oligonucleotide.
Identical results were seen with a radiolabeled oligonucleotide containing WT1 binding site 2. A single species was found in the nuclear extract from the control cells, whereas two species, one of which migrated more rapidly than the species from the control cells, were found in the extracts from the WT1-expressing cells. Both species were inhibited by an excess of the unlabeled oligonucleotide, but only the more slowly migrating species was inhibited by an oligonucleotide containing a mutation that abrogates binding by WT1 (Fig. 3C).
Electrophoretic mobility shift assay was also carried out using these nuclear extracts and an oligonucleotide containing the third putative WT1 binding site, this one located in the 5Ј untranslated portion of the mRNA. Unlike the other two sites, there was no activity in the nuclear extract from the control 32DV4 cells capable of retarding the migration of the oligonucleotide containing binding site 3 (Fig. 3D). An activity capable of binding this oligonucleotide was present in the nuclear extract from WT1(Ϫ/Ϫ)-expressing cells (Fig. 3D). This activity was inhibited by an excess of unlabeled oligonucleotide but not by an excess of unlabeled oligonucleotide containing a mutation that abrogates binding by WT1. Thus, this sequence is also bound by WT1(Ϫ/Ϫ), but unlike the other two binding sites, it is not bound by an activity present in the untransfected cells.
The WT1 Binding Site in Exon 1 Is Necessary for Cyclin E Promoter Repression-Having demonstrated that WT1(Ϫ/Ϫ) can bind to each of the putative binding sites in the cyclin E promoter, we wanted to determine the relative importance of FIG. 3. WT1(؊/؊) binds to the putative binding sites in the cyclin E promoter. A, nuclear extracts from 32DV4 and 32DWT3 cells were subjected to Western blotting analysis with an antibody that recognizes both murine and human WT1. Nuclear extracts were incubated with radiolabeled oligonucleotides alone or in the presence of an excess of unlabeled oligonucleotide (Cold Oligo) or unlabeled mutant oligonucleotide (Mutant Oligo). Protein-DNA complexes were separated by PAGE followed by autoradiography. This experiment was performed with oligonucleotides containing the sequence surrounding binding site 1 (B), site 2 (C), and site 3 (D). In each case, the arrows indicate the WT1-oligonucleotide complex and the nonspecific band. each of these sites for the regulation of the promoter. We therefore generated deletion mutants of the cyclin E promoter lacking each binding site alone and in combination. Luciferase reporter constructs containing these deletion mutants were cotransfected into NIH 3T3 cells with the plasmid encoding WT1(Ϫ/Ϫ). Similar to what was seen with the full-length promoter, deletion mutants lacking site 1, site 2, or both of these sites, were strongly repressed by WT1(Ϫ/Ϫ). In contrast, deletion mutants lacking site 3 alone, or in combination with the other binding sites, were no longer repressed by this isoform of WT1 (Fig. 4). Although the basal activity of the constructs lacking binding site 3 was significantly less than the other constructs (which might reflect an effect on either RNA stability or on the basal rate of transcription), there was still sufficient activity present compared with the promoterless pGL2basic (Fig. 4) that an effect of WT1(Ϫ/Ϫ) could have been observed. This finding suggests that transcriptional repression of the cyclin E promoter requires the presence of a WT1 binding site downstream of the transcription start site.

DISCUSSION
Although WT1 was originally identified as a tumor suppressor gene, recent literature supports the view that it may have oncogenic potential in acute leukemia and in breast cancer (16,17). The protein product of the WT1 gene has transcriptional regulatory activity. A number of target genes have been identified that might be important for neoplasia. Because induction of WT1 expression leads to up-regulation of p21, we examined the promoters of other genes that function in the G 1 /S transition for the presence of the well defined consensus DNA binding sequence for the WT1(Ϫ/Ϫ) isoform. We discovered that the promoter for cyclin E, another key regulator of cell cycle progression, contains three potential WT1(Ϫ/Ϫ) binding sites. Cyclin E functions by binding to and activating Cdk2 (18,19). The cyclin E-Cdk2 complex is a critical promoter of initiation of the S phase of the cell cycle.
We have demonstrated here that WT1(Ϫ/Ϫ) binds to all three consensus binding sites in the cyclin E promoter, nega-tively regulates the promoter in transient transfection assays, and in intact cells, down-regulates cyclin E protein levels. These findings support the idea that cyclin E is a direct target of the transcriptional regulatory activity of WT1. This conclusion is strengthened by our finding that deleting the binding site from the 5Ј untranslated region of exon 1 eliminates the ability of WT1 to repress the promoter in transient transfection assays. Deletion of the region surrounding this site clearly decreases the basal level of promoter activity substantially; nevertheless, there is still significant residual activity, which WT1(Ϫ/Ϫ) does not suppress.
Our finding that WT1(Ϫ/Ϫ) directly suppresses cyclin E complements the published observation that induction of WT1(Ϫ/Ϫ) in Saos2 cells up-regulates the expression of p21 WAF1/CIP1 (11). Those studies were limited in that no evidence, such as promoter reporter assays or electrophoretic mobility shift assays, was presented to support the conclusion that WT1(Ϫ/Ϫ) directly affects the p21 promoter, leaving open the possibility that some other WT1 target is a p21 transcriptional activator. The present report, therefore, contains the first demonstration that the WT1(Ϫ/Ϫ) isoform is a direct transcriptional regulator of a key gene involved in cell cycle progression.
Interestingly, in our 32D cl3 cells stably transfected with an inducible WT1(Ϫ/Ϫ) construct, we saw no evidence that induction of WT1(Ϫ/Ϫ) alters p21 expression (Fig. 1). There are two possible explanations for this observation. One possibility is that WT1 is not the direct transcriptional activator of p21. In this case, the WT1 target that mediates the up-regulation of p21 may not be expressed in hematopoietic cells. Alternatively, this may represent a cell type-specific effect of WT1 on the p21 promoter. There is precedence for this with bcl-2. In transient transfection assays, WT1(Ϫ/Ϫ) activates the bcl-2 promoter in Saos2 and CV-1 cells but represses the identical reporter construct in HeLa cells (12).
Cyclin E and p21 are both key regulators of progression from G 1 into S phase. Cyclin E is a positive regulator of this process, binding to and activating cyclin-dependent kinase 2 (Cdk2). FIG. 4. Deletion mutagenesis of the cyclin E promoter. NIH 3T3 cells were cotransfected with the indicated reporter construct and a WT1(Ϫ/Ϫ) expression vector. The left panel shows a schematic representation of the various reporter constructs, with the regions surrounding the putative WT1 binding sites in gray. pGL2basic, which lacks a promoter, was included as a negative control because of the relatively low luciferase activity of the constructs lacking binding site 3. Relative luciferase activity was normalized to cotransfected Renilla luciferase. Error bars represent standard error of the mean as determined in triplicate assays. Activated Cdk2 phosphorylates Rb, which is necessary for progression from G 1 into the S phase of the cell cycle (19). One of the roles of p21, on the other hand, is to inhibit the activity of Cdk2 (20). There have been other reports in the literature suggesting that the G 1 /S phase transition is regulated, at least in part, by WT1 (21). Evidence that WT1(Ϫ/Ϫ) up-regulates an inhibitor of G 1 /S progression (p21) and down-regulates a promoter of cell cycle progression (cyclin E) strongly suggests that inhibition of this point of the cell cycle is a critical role for WT1.
How does the finding that WT1(Ϫ/Ϫ) is a negative regulator of cell cycle progression fit with a putative oncogenic role? WT1 is overexpressed in the majority of cases of acute leukemia (16), and mutant forms of WT1 encoding nonsense mutations predicted to encode truncated proteins lacking the DNA binding domain have been cloned from the leukemic blasts of several patients (13). Similar mutations have been identified in patients with Denys-Drash syndrome, and these mutants have dominant negative activity in vitro (8). WT1 functions as a dimer, and both homo-and heterodimers can form (15). In leukemias that overexpress wild type WT1, the most abundant isoform is WT1(ϩ/ϩ), which not only suppresses the cyclin E promoter poorly as a homodimer but also inhibits the ability of WT1(Ϫ/Ϫ) to repress this promoter (presumably through heterodimerization). Overexpression of this isoform could "derepress" cyclin E transcription, resulting in inappropriate overexpression of this target. In cases of acute myeloid leukemia involving a truncated WT1, this dominant negative form of the molecule should also inhibit the activity of full-length WT1(Ϫ/Ϫ) toward target promoters such as cyclin E through heterodimer formation. In fact, cyclin E is reported to be overexpressed in the majority of cases of acute myeloid leukemia (22), suggesting that this may be a major mechanism contributing to leukemogenesis. Furthermore, we have found that induction of WT1(Ϫ/Ϫ) in the stably transfected cell line described in this paper potentiates granulocyte-colony stimulating factor-mediated differentiation, 1 whereas another group reported that constitutive expression of WT1(ϩ/ϩ) in 32D cl3 cells interferes with differentiation (23).
The regulation of cyclin E protein levels has recently been a prominent focus of cancer researchers. Cyclin E levels vary with the cell cycle, with low levels present in early G 1 , rising through G 1 to peak at the G 1 /S transition, falling rapidly during early S phase, and remaining low through G 2 , M, and early G 1 phases. This precise regulation of the protein is achieved by balancing transcription and proteasome-dependent degradation. Dysregulation of cyclin E is associated with premature entry into S phase, genomic instability, and tumorigenesis. An F-box protein, which mediates association with ubiquitin ligase, specific for cyclin E has recently been described (24 -26), and mutations in the gene encoding this protein, which would lead to elevated cyclin E levels, have been implicated in breast and ovarian cancer (25,26). Our discovery that WT1(Ϫ/Ϫ) is a transcriptional repressor of cyclin E suggests a complementary mechanism by which cyclin E might be aberrantly overexpressed in tumors through derepression of the cyclin E pro-moter by either overexpression of the WT1(ϩ/ϩ) isoform or by expression of a dominant-negative WT1 mutant. Experiments to further define the interactions between the different isoforms of WT1 in regulating the cyclin E promoter are ongoing. WT1 has been implicated in a variety of solid tumors other than Wilms' tumor. For example, WT1, predominantly the (ϩ/ϩ) isoform, is overexpressed in a significant proportion of breast cancers (17). The ability of this isoform to derepress the cyclin E promoter might be a contributory factor in mammary carcinogenesis. Whether cyclin E derepression is involved in other WT1-associated tumors such as malignant mesothelioma remains to be determined.