H19 mRNA-like Noncoding RNA Promotes Breast Cancer Cell Proliferation through Positive Control by E2F1*

The imprinted H19 gene has riboregulatory functions. We show here that H19 transcription is up-regulated during the S-phase of growth-stimulated cells and that the H19 promoter is activated by E2F1 in breast cancer cells. H19 repression by pRb and E2F6 confirms the E2F1-dependent control of the H19 promoter. Consistently, we demonstrate by chromatin immunoprecipitation assays that endogenous E2F1 is recruited to the H19 promoter in vivo. The functionality of E2F promoter sites was further confirmed by gel shift and mutagenesis experiments, revealing that these sites are required for binding and promoter response to E2F1 exogenous expression and serum stimulation. Furthermore, we show that H19 overexpression confers a growth advantage on breast cancer cells released from growth arrest as well as in asynchronously growing cells. The H19 knockdown by small interfering RNA duplexes impedes S-phase entry in both wild-type and stably H19-transfected cells. Based on these findings, we conclude that the H19 RNA is actively linked to E2F1 to promote cell cycle progression of breast cancer cells. This clearly supports the H19 oncogenic function in breast tumor genesis.

The H19 gene is one of the first genes to have been proven to be imprinted. It is located on chromosome 11 p 15.5 and lies within 200 kbp downstream of the IGF-2 gene (1). These two genes are imprinted in opposite directions, so that the paternal IGF-2 and the maternal H19 alleles are selectively expressed (2,3). H19 encodes a spliced and polyadenylated RNA that lacks conserved open reading frames but does have a conserved secondary RNA structure (4). Extensive deletions and/or point mutations in the 5Ј-long untranslated region of an ectopic human H19 RNA enable 26-kDa protein translation (5), but no endogenous translation product has so far been identified (6,7). Therefore, it was quickly proposed that H19 RNA functions as a riboregulator (8).
H19 expression is developmentally regulated. It is abundantly expressed in both extraembryonic and fetal tissues and is repressed after birth except in a few adult organs, particularly in the mammary gland (9,10). Since the first mention of H19 in 1984 by Pachnis et al. (6), its functions have only begun to emerge. It has been reported that H19 RNA was involved in the repression of the IGF-2 oncogene by affecting its transcription (11) or its translation (12). Recently, we brought evidence that the H19 gene post-transcriptionally up-regulates the thioredoxin level, a key protein of the cellular redox metabolism (13).
Despite interesting new insights, the status of the H19 gene in cancer is still a matter of debate. It has been suggested that H19 functions as a tumor suppressor in some Wilms' tumors, embryonic rhabdomyosarcoma, and the Beckwith-Wiedmann cancer predisposing syndrome (14 -16). Consistently with this function, some studies conclude that it down-regulates the IGF-2 factor (11). By contrast, other studies including ours suggest that H19 may play a key role in tumorigenesis and could match the cell aggressiveness (17,18).
H19 activation has also been reported in various cancer tissues: breast (9,10,19), bladder (20,21), lung (22), and esophageal cancers (23). Its oncogenic role has been well documented in the bladder, since it is considered as an oncodevelopmental marker (24) and regulates genes involved in metastasis and blood vessel development (25). These observations support a H19 role in tumor invasion and angiogenesis. In breast cancer, the oncogenic role of H19 has been well established (26), but the precise gene function in cellular processes is not yet understood. Furthermore, H19 promoter regulation remains widely undetermined. Only a few studies have reported H19 expression to be modulated by cytokines in cells (27)(28)(29)(30). However, despite putative regulatory sequences localized in the H19 promoter and the identification of potential transacting factors, no direct regulation of the H19 promoter has so far been described. In a previous work, we reported a negative regulation of H19 promoter activity by the tumor suppressor p53 that would be mediated by protein/protein interactions (31). Consistently, Sorensen et al. (32) proposed a functional interaction between p53 and DP1 (an E2F1 transcriptional partner) to explain the loss of the dihydrofolate reductase promoter activation by E2F1 in SAOS-2 cells. Because of the presence of two putative E2F consensus sites in the H19 minimal promoter, we hypothesized a similar mechanism for H19 promoter control, and we anticipated a role for E2F1 in the regulation of H19 promoter activity. Consequently, we have further investigated the H19 physiological function through this regulation, and we have logically focused on cell cycle progression.
E2F transcription factors consist of a related protein family that includes seven distinct E2F members, which act on gene promoters regulated during the cell cycle (33,34). The first member of the E2F family to be cloned, E2F1, is considered to be the critical factor in G 1 /S transition (35). E2F1 has been shown to be involved in the transcriptional regulation of several genes, whose products participate in cell cycle progression and DNA synthesis and whose expression is up-regulated at the cell cycle G 1 /S transition (36 -38). We focused our work on E2F1 protein, which does have biological consequences in carcinogenesis. Indeed, E2F1 enhances neoplasia in the skin of transgenic mice and causes tumor formation in the liver (39,40). Furthermore, E2F1 is involved in breast cancer development through transcriptional activation of the breast cancer susceptibility gene BRCA1 (41) and serves as the primary link between proliferation control and apoptosis (42).
The purpose of the present study was thus to investigate the role of H19 in the cell cycle via a regulation by E2F1. We report that the cell cycle-dependent regulation of the H19 gene is primarily controlled by E2F1 through binding to H19 promoter. pRb and E2F6, two E2F-dependent transcription inhibitors, repress H19 gene expression. We further focused our study on the biological effect of the H19 gene on cell cycle progression and particularly at the G 1 -S transition. We show that H19 overexpression confers an obvious growth advantage in breast cancer cells released from growth arrest as well as in asynchronously growing cells. In addition, the knock-down of H19 expression by small interfering RNA (siRNA) impedes progression through the S-phase of the cell cycle. In conclusion, we elucidate the involvement of the H19 mRNA-like noncoding RNA in cancer cell proliferation. Our results clearly shed light on the oncogenic status of H19 in breast cancer cells.

MATERIALS AND METHODS
Cell Culture and Breast Tissues-The human breast cancer epithelial cell lines were obtained from the ATCC and maintained routinely in MEM, 1 containing 5% fetal calf serum (FCS). MCF-7 and T47D are estrogen-sensitive cells. BT20 and MDA-MB-231 are estrogen-insensitive cell lines. Normal breast epithelial cells come from primary culture of normal breast tissue resections obtained from modeling surgery. The highly tumorigenic MCF-7Ras cell line corresponds to the H-Ras-transfected MCF-7 cell line. Human embryonic kidney cells (HEK 293) provided by Dr. D'Halluin (Institut de Recherche sur le Cancer de Lille, France) were grown in Dulbecco's modified Eagle's medium supplemented with 10% FCS. To bring cells to quiescence, cultures at 30% confluence were washed with MEM and then serum-starved for 24 h. Cells were serum-stimulated by the addition of 10% FCS to the medium. All cell lines were incubated at 37°C in a humidified atmosphere with 5% CO 2 and 95% air. Breast cancer tissues came from biopsies corresponding to intraductal cancers.
Growth Assays-Mock-transfected and H19-stably transfected MDA-MB-231 cells were previously cloned (26). For growth assays of serumstimulated cells, cells were plated in 6-well plates at 100,000 cells/well, serum-starved for 24 h, and induced to reenter the cell cycle by the addition of 10% FCS medium for 48 h. For RNA interference assays, cells were plated in 6-well plates at 150,000 cells/well on day 1. The day after, cells were transfected with respective siRNA sequences. Then cells were grown in 5% FCS containing medium until day 5. Cells were counted each day in a Malassez plate. Student's t test was performed, and the p values obtained are indicated in the figures by asterisks with the following meanings: *, p values Ͻ 0.05 considered to be statistically significant; **, p values Ͻ 0.01 considered to be highly statistically significant.
Reporter Plasmids and Expression Vectors-pCMV-E2F1 was generously provided by Prof. Helin (Institute of Oncology, Milan, Italy). pCMV-E2F1⌬ is an E2F1 mutant deleted in the transactivation domain. PSG5-Rb was a gift from Dr. Bégue (Institut de Biologie de Lille, France). PHSCdc6-Luc was provided by Prof. Othani (Human Gene Sciences Centre, Tokyo, Japan). The H19-luciferase reporter gene corresponds to the plasmid pGL2 carrying the H19 minimal promoter region (823 bp), associated with the luciferase gene reporter described by Dugimont et al. (31). The H19 mutant reporter plasmids were constructed by oligonucleotide-directed mutagenesis using PCR. To obtain the single mutant named hereafter Mut-I, we carried out two similar PCRs with the following primers (mutations are indicated in boldface type): Z1, 5Ј-AACAACCCTCACCAAAGGCC-3Ј; Y1, 5Ј-AGGGGGTTT-GAAGCACTTCC-3Ј; Y2, 5Ј-AGGGCAGGGGCGGGAATTCTGGAAG-GCCA-3Ј; Z2, 5Ј-TCCTCGGTCCTAGCCCGG-3Ј. A final PCR with primers Z1 and Z2 permitted us to obtain the total length of the mutated promoter. To obtain the single mutant Mut-II, we carried out a PCR with Z1 and Z3 5Ј-CTCGGTCCTAGCCCGGGCTTTTTCTA-ACTGGGGTGGCCTTCCAGAAT-3Ј. The double mutant (DMut) was obtained by a PCR with Z1 and Z3 using the Mut-I PCR product as matrix. All of the PCR products were cloned into the pCR2-1 TOPO vector (Invitrogen). Recombined vectors were selected by enzymatic digestion, and mutations were verified by sequencing. The mutated promoters were cloned into pGL2 plasmid associated with luciferase reporter gene.
Transfection Assays-Cotransfections were performed by using Lipofectin TM reagent (Invitrogen). 1.2 g of the reporter plasmid, 0.5 g of the expression plasmid, and 0.2 g of ␤-galactosidase were used for 150,000 cells/well, seeded 2 days before transfection. Cells were incubated with transfection reagent for 6 h at 37°C, followed by a change to a culture medium containing 10% FCS. Cells were harvested with a reporter lysis buffer (Applied Biosystems) 24 h (for MCF-7 cells) and 48 h (for HEK cells) after transfection. Luciferase and ␤-galactosidase activities were measured by using an AB luciferase assay kit and a ␤-galactosidase assay system kit, as specified by the manufacturer, with a Lumat LB 9501 luminometer (Berthold).
RNA Interference-RNA interference was carried out by using synthetic siRNA duplexes, as described by Elbashir et al. (43). Two synthetic siRNA duplexes (siH19A and siH19B), corresponding to the H19 mRNA sequence 5Ј-CCCACAACAUGAAAGAAAU-3Ј and 5Ј-GCUA-GAGGAACCAGACCUU-3Ј, respectively, were used to inhibit H19 RNA expression. A synthetic siRNA duplex (siE2F) corresponding to a previously described E2F1 mRNA sequence 5Ј-GGGAGAAGUCACGC-UAUGA-3Ј was used to inhibit E2F1 protein expression (44). A synthetic siRNA duplex (siGFP) corresponding to the green fluorescent protein mRNA sequence 5Ј-GCUGACCCUGAAGUUCAUC-3Ј was used as a negative control. The siRNA duplexes were purchased from Eurogentec. Cells were grown on coverslips in 6-well plates and transfected with 400 pmol of siRNAs using Jetsi (Eurogentec, Angers, France), as recommended by the manufacturer. To monitor the transfection efficacy, a tagged siRNA duplex was transfected in parallel, and the transfection rate was evaluated by FACS and seen to correspond to 80 -90% of transfected cells. After transfection, cells were lysed for total RNA isolation or counted for growth assays.
Northern Blot Analysis-Total RNA was extracted using the guanidium isothiocyanate-CsCl gradient method (45). RNA (20 g) was denatured, electrophoresed through a 1.2% agarose gel containing formaldehyde, and transferred by capillarity onto a nitrocellulose membrane (Hybond-C-extra; Amersham Biosciences). After being baked for 2 h at 80°C, the membrane was hybridized at 42°C with [␣-32 P]dCTPlabeled random primed cDNA probes (Megaprime Labeling System; Amersham Biosciences). The H19 cDNA probe was a 1.3-kbp PstIdigested fragment containing the end of the first exon and exons 2-5. An acid ribosomic phosphoprotein 0 cDNA probe was also used to normalize the H19 signal.
ACA-3Ј and 5Ј-TGGCCATGAAGATGGAGTCG-3Ј for H19 near the si-H19B; 5Ј-CCGTGGACTCTTCGGAGAAC-3Ј and 5Ј-GGGACAACAGC-GGTTCTTGC-3Ј for E2F1, and 5Ј-GTGATGTGCAGCTGATCAAGAC-T-3Ј and 5Ј-GATGACCAGCCCAAAGGAGA-3Ј for RPLPO (human acidic ribosomal phosphoprotein PO), which was used as a reference gene. The subsequent PCR conditions were 40 cycles, carried out in the following manner: 95°C for 15s, 60°C for 20s, and 72°C for 30s. Data were analyzed using the MX4000 PCR system software (Stratagene, Amsterdam, The Netherlands) with the SYBRGreen option (with dissociation curves). Standard curves were performed on serial dilutions of a PCR product for H19 and on serial dilutions of genomic human DNA for RPLPO. Values were obtained by the calculation methodology recommended by Pfaffl et al. (46): ratio ϭ (cycle number Ϫ b/a) target/(cycle number Ϫ b/a) reference (where a ϭ slope of the standard curve and b ϭ ordinate of origin).
Western Blot-Nontransfected or RNAi-transfected cells were harvested in lysis buffer. 30 g of lysates were subjected to SDS-PAGE, transferred onto a nitrocellulose membrane (Immobilon-P; Millipore Corp.) by electroblotting (30 V, overnight), and probed with anti-E2F1 (sc-251; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or anti-actin antibodies. After washing, the membranes were incubated with horse peroxidase-conjugated anti-mouse immunoglobulin (Jackson Laboratories). The reaction was revealed using the chemiluminescence kit ECL (Amersham Biosciences) with Eastman Kodak Co. hyperfilm.
Fluorescent Analysis and Cell Sorting (FACS)-Cells from a 10-cm diameter dish were trypsinized, centrifuged, washed with PBS, and fixed in 70% ethanol. For transfected cells, selection was made with the cotransfection of a GFP expression plasmid. In this case, cells were previously fixed in paraformaldheyde solution for 30 min at 4°C, washed twice with PBS, and subsequently fixed in 70% ethanol for 30 min at 4°C. For propidium iodine (PI) staining, cells were washed twice in PBS, centrifuged, and resuspended in 500 l of PI buffer (500 l of PBS, 100 g of PI, and 100 g of RNase A). After 30 min of incubation at 37°C, the samples were analyzed with a Becton Dickinson FACScan.
Chromatin Immunoprecipitation Assays-Detection of promoterbound E2F1 proteins was assessed by chromatin immunoprecipitation assays, essentially as previously described (47,48). Cells were treated with formaldehyde to form cross-links between E2F and associated promoter regions. Chromatin was then isolated, fragmented by sonication, and subjected to immunoprecipitation by using antibody directed against E2F1 (catalog number SC-193; Santa Cruz Biotechnology). An antibody against the FLAG epitope (catalog number SC-807; Santa Cruz Biotechnology) was used as a negative control (nonrelevant antibody). To detect the H19 gene in protein-DNA complexes, a 141-bp fragment located in the promoter was amplified by PCR using oligonucleotides 5Ј-GTCTGGGAGGGAGAGGTCCT-3Ј and 5Ј-CCACTCTCT-CTGCACACGAC-3Ј. The primers used to detect the E2F1 gene promoter were the same used by others (48). The two genes were detected by 34 cycles of PCR. A negative control primer pair located downstream to the enhancers at ϩ40510 bp to the H19 start site was used to attest to the binding specificity. The primer sequences were as follows: 5Ј-CTCTCCTCTCTCGACCTGTC-3Ј and 5Ј-GCAGCAGTGTCTCAGGA-GAG-3Ј. For this condition, PCR was extended to 42 cycles to ascertain the irrelevance of amplifications obtained for immunoprecipitations with nonrelevant and ␣-E2F1 antibodies.
Electrophoretic Mobility Shift Assay-Double-stranded oligonucleotides were generated corresponding to the wild-type or E2F-mutated site II of the H19 promoter. The wild-type oligonucleotide was end-labeled for use as a probe. The E2F binding site is underlined, and the mutated nucleotides are shown in boldface type: H19WT, 5Ј-GAATTCTGGCG-GGCCACCCCA-3Ј; H19mut, 5Ј-GAATTCTGGAAGGCCACCCCA3Ј. These double-stranded oligonucleotides were also used as competitors at a 400-fold molar excess. As a positive control, we also competed with a typical E2F site from Nushift TM Kit (Geneka Biotechnology, Inc., Strasbourg, France). For binding experiments, the following components were mixed and preincubated at room temperature for 15 min and depending on the experiment: 4 l of complete cell extracts of COS7 cells transiently both transfected and untransfected by E2F1 vector, 1 g of competitors, 4 l of E2F1 monoclonal antibody (Nushift TM kit), and 10 l of binding buffer (500 ng of sheared salmon sperm DNA, 25 mM Hepes, pH 7.6, 25 mM KCl, 1 mM EGTA, 2 mM MgCl 2 , 0.05% Nonidet P-40, 10% glycerol) in a final assay volume of 20 l. Following preincubation, the labeled oligonucleotide was added (2.5 ng), and the mixtures were incubated for another 15 min at room temperature. Samples were loaded onto a 4% nondenaturating polyacrylamide gel in 1ϫ TBE (22.5 mM Tris borate, 0.5 mM EDTA) at 4°C. The gel was dried, and autoradiography was performed.

H19 Expression Is Transcriptionally
Up-regulated during the S-phase of Cells Stimulated to Enter the Cell Cycle-Cell growthdependent control of the E2F1 gene and E2F1-regulated genes has been described (37,49,50). We previously showed downregulation of H19 by p53 and anticipated a possible role for p53 in E2F1 activity repression (31). These indications led us to examine whether H19 expression was cell cycle-dependent in MCF-7 cells. Cells were serum-starved and subsequently induced to reenter the cell cycle by serum addition. Cells were harvested after starvation or at various times following serum addition. Fig. 1A shows the FACS analysis of these cells, indicating that most of them are in the S-phase after 32 h of stimulation. Cells treated under these various conditions were checked for expression levels of both H19 and E2F1 RNAs. Expression levels of both genes were low in serum-starved cells (Fig. 1B). As expected, E2F1 transcription was up-regulated by serum addition, with a peak at the G 1 /S-phase boundary (16 -20 h). In parallel, H19 mRNA synthesis was activated and reached a high expression level at 32 h, coinciding with the greatest number of cells in S-phase. Thus, both genes are transcriptionally activated after serum addition, although the timing differs, since the target H19 gene response takes longer than the E2F1 induction. Collectively, these results show the cell cycle-regulated expression of the H19 gene and suggest that H19 is likely to be regulated by E2F1.
E2F1 Activates the H19 Promoter-The H19 gene possesses a TATA-less promoter, a feature reminiscent of many known E2F-responsive genes. Moreover, a screening of the H19 promoter sequences revealed two putative E2F recognition sites in proximity to the H19 transcription start site (sites I and II boxed in Fig. 4A). These properties, together with the finding that H19 expression was cell cycle-regulated, prompted us to examine whether the H19 promoter was directly regulated by E2F1. We tested this hypothesis by performing transient transfections with a luciferase reporter vector under the control of the minimal H19 promoter (31). This reporter vector was cotransfected into MCF-7 cells together with a plasmid-expressing E2F1 wild-type protein (E2F1), or an E2F1 version deleted in its transactivation domain (E2F1⌬). E2F1 expression in FIG. 2. H19 promoter activation by E2F1. A, MCF-7 cells were transfected with 1.2 g of H19-luciferase reporter gene and 0.5 g of pCMV-E2F1 or pCMV-E2F1⌬ expression plasmids. Parental pCMV vector was used as controls. The Cdc6-luciferase gene was used as a positive functional control. Transfected cells were cultured in medium with 5% FCS for 24 h. B, inhibition of E2F1 expression by specific siRNAs. MCF-7 cells were transfected with 400 pmol of siE2F1 or siGFP used as control conditions, and proteins were processed for Western blotting using an anti-E2F1 antibody. C, knock-down of E2F1 decreases H19 expression. MCF-7 cells were transfected with the E2F1-specific siRNA (siE2F1) or with the control siRNA targeting the green fluorescent protein (siGFP). Endogenous H19 expression was measured by real time PCR and expressed in relative H19 expression with regard to the RPLPO reference gene, as described under "Materials and Methods." E2F6 (D) and pRb (E) act as negative regulators of H19 promoter. MCF-7 cells were transfected with 1.2 g of H19-luciferase reporter gene and 0.5 g of expression vectors (pCMV-E2F6 and pSG5-Rb and/or pCMV-E2F1). Empty pCMV and pSG5 vectors were used as controls. Transfected cells were cultured in medium containing 5% FCS for 24 h. For all relative luciferase activities, pCMV-␤-galactosidase (0.2 g) was cotransfected as an internal control. Luciferase and ␤-galactosidase activities of cell extracts were measured. Luciferase activity of cell extracts was normalized to ␤-galactosidase activity and is plotted for each transfected empty vector. All values shown represent the mean values of three independent experiments. MCF-7 cells induced a 5-fold activation of the H19 promoter 24 h after transfection, whereas E2F1⌬ severely reduced H19 promoter activity (about 25% remaining activity compared with the control basal level) (Fig. 2A). The latter feature was probably due to the competition of E2F1⌬, considered to be a negative transdominant mutant, with the endogenous E2F1 wild-type proteins. We obtained similar results in two nonrelated cell lines (HeLa and HEK-293 cells) treated following the same procedure (data not shown). In parallel, as a functional control in MCF-7 cells, we checked E2F1 plasmid activity by using a known E2F1-regulated gene (i.e. the Cdc6 gene) (49). As expected, this gene was obviously up-regulated by E2F1, the phenomenon reaching about a 12-fold increase.
Then we investigated the effect of E2F1 inhibition by RNA interference on the endogenous H19 expression in MCF-7 cells. When the E2F1 protein level is decreased by siRNA targeting human E2F1 mRNA (Fig. 2B), H19 expression measured by real time RT-PCR is strongly reduced with regard to the negative control siRNA targeting green fluorescent protein (siGFP) (Fig.   2C). Finally, we tested the effect of two E2F1-negative regulators: E2F6, which lacks the transactivation domain (51,52), and pRb, which complexes and inactivates E2F1. As expected, E2F6 repressed the H19 basal expression by about 50% in transiently transfected MCF-7 cells (Fig. 2D). Results shown in Fig. 2E indicate that pRb had only a weak direct negative effect on H19 promoter activity in these cells but was able to completely suppress the H19 transactivation by E2F1. This suggests that pRb indirectly represses the H19 promoter through the inhibition of E2F1 activity. Taken together, these results confirm the E2F1 involvement in H19 promoter regulation.
In Vitro and in Vivo Association of E2F Proteins with the H19 Gene-As mentioned above, we found two sequences in the H19 promoter (I and II) matching the consensus of an E2F binding site. The functionality of the E2F sites was investigated by electrophoretic mobility shift assays using wild-type and mutated probes. We chose a radiolabeled probe, which was a sequence of 21 nucleotides overlapping the site II (Fig. 3A). This probe was able to generate complexes with nuclear ex -FIG. 3. In vitro and in vivo E2F1 binding to the H19 promoter. A, nucleotide sequences of the WT or mutated H19 probes used in gel mobility shift assays. B, gel mobility shift assays were performed with COS-7 cell extracts and the end-labeled E2F fragment from the H19 promoter containing the wild-type site II as a probe. Lane 1, the probe only (probe); lane 2, the probe and the nontransfected cell extracts (NTCE); lane 3, the probe and the E2F1-transfected cell extracts (TCE). Competition experiments were performed by adding a 400-fold molar excess of unlabeled oligonucleotides to the gel shift reaction mixtures. The competitor oligonucleotides used were the wild-type site II (lane 4) and a mutant form of this sequence (lane 5). As a positive control, we also competed with a typical E2F site from the Nushift TM kit (lane 6) and a mutant form of this sequence (lane 7). A supershift was obtained when an antibody directed against E2F1 was added to cell extract (lane 8). C, immunoprecipitation of E2F-associated H19 promoter fragment. HeLa cells were treated with formaldehyde to create cross-links between transcription factors and chromatin. The chromatin was isolated, sheared, and immunoprecipitated using antibody directed against E2F1 or nonrelevant antibody (NR Ab). The presence of chromatin fragments corresponding to the H19 gene or to the E2F1 gene promoter was assessed by semiquantitative PCR using gene-specific primers. Recovery of H19 and E2F1 gene fragments from the protein-DNA extracts (prior to the immunoprecipitation) is shown in lane 1. Negative control primers correspond to primers located out of the H19 promoter (at ϩ40,510 bp to the H19 transcription start site). Weak amplifications, appearing only after 42 cycles of PCR in immunoprecipitations with nonrelevant and ␣-E2F1 antibodies, are nonrelevant. The PCRs were separated by electrophoresis on a 2% agarose gel. tracts of E2F1/DP1 transfected COS-7 cells (Fig. 3B, lane 3, TCE) but was not able to do so in nontransfected cells (Fig. 3B, lane 2, NTCE). Complex formation was efficiently disrupted in the presence of excess quantities of the same nonlabeled oligonucleotide (Fig. 3B, lane 4), as well as with an excess of an oligonucleotide containing a typical E2F binding site (Fig. 3B,  lane 6). By contrast, the same oligonucleotides mutated in a base pair crucial to the E2F1 binding (50) did not compete with the wild-type probe (Fig. 3B, lanes 5 and 7). To confirm the binding specificity, we incubated the nuclear extracts with an antibody raised against E2F1. This led to a supershift of the complex (Fig. 3B, lane 8). Thus, E2F1 can be seen to bind directly to the H19 promoter throughout the site tested.
In a more physiological context, using the in situ H19 promoter, we tested if E2F proteins associate with these regulatory sequences in vivo by chromatin immunoprecipitation experiments (Fig. 3C). Chromatin was subjected to immunoprecipitation by using antibody directed against E2F1. The presence of the H19 gene promoter was detected by amplifying a promoter region located between the two E2F sites. A DNA fragment corresponding to the E2F1 gene promoter was amplified as a positive control. As shown in lane 3, a genomic DNA fragment containing the H19 promoter co-immunoprecipitated with E2F1. Both E2F1 and H19 genes were efficiently detected by PCR when the input chromatin (after fragmentation but before immunoprecipitation) was subjected to PCR amplification (lane 1). Negative control for the immunoprecipitation consisted of performing the immunoprecipitation with an irrelevant antibody (lane 2). Negative control primers correspond to oligonucleotides located out of the H19 promoter. Weak amplifications in lanes 2 and 3 were obtained only when the PCR was extended to 42 cycles and correspond to the experimental background.
E2F Recognition Sites Are Required for the Activation of the H19 Promoter by E2F1 Proteins-We investigated the role of the E2F binding sites in functional assays. To directly test the contribution of each of the two potential E2F sites (Fig. 4A), we generated point mutations by oligonucleotide-directed mutagenesis of these sites in the H19-luc reporter construct (Fig.  4B), either individually (Mut-I or Mut-II) or in combination (double mutation; DMut). The mutated nucleotides have been described to be essential for E2F binding and activity (38,50). MCF-7 cells were cotransfected with these reporter plasmids and with either an empty vector (pCMV) or with the pCMV-E2F1 expression vector. Mutation of either one or both E2F sites almost completely abolished the H19 promoter activation by ectopic E2F1 proteins, indicating that both E2F sites actually contribute to the H19 promoter regulation (Fig. 4C). We also considered another cell model constitutively expressing high levels of active E2F1 protein. The human embryonic kidney cell line 293 (HEK-293) has an incorporated functional E1A protein derived from adenovirus 5, which constitutively releases free and active E2F in high quantities (53-55). We performed single transfections in these cells with wild-type or mutated E2F site reporter vectors in order to evaluate the H19 promoter activation level by endogenous E2F proteins. Mutations strongly reduced the promoter basal activity (Fig. 4D). These data demonstrate that E2F1 directly regulates the H19 promoter through its two E2F recognition sites.
E2F Sites Mediate H19 Promoter Activation during S-phase-It is likely that E2F1 regulates H19 expression via the E2F sites in its promoter during growth stimulation. To test this possibility, we examined the wild-type and E2F sitemutated H19 promoter activities in serum-starved MCF-7 cells, as well as in cells stimulated to reenter the cell cycle (Fig.  5A). The cells were submitted in parallel to a FACS analysis (Fig. 5B). The H19 promoter activity was low in serum-starved cells (0 h) and was markedly up-regulated when cells reached the S-phase (28 h). This observation is consistent with our analysis of endogenous H19 expression (Fig. 1). Mutations of the two E2F sites (DMut) abolished the promoter activation by serum stimulation (Fig. 5A). These results demonstrate that the H19 promoter E2F binding sites efficiently contribute to H19 transcriptional activation during growth stimulation.
Relationship between E2F1 and H19 Expressions in Breast Epithelial Cells-We studied the expression levels of E2F1 mRNA and E2F1 proteins (Fig. 6, A and B) and H19 RNA (Fig.   6C) in several breast epithelial cells. Normal breast epithelial cells were used as a control. Other cell lines are either estrogendependent (T47D, MCF-7) or estrogen-independent (BT20, MDA-MB-231) cancer cell lines. MCF-7 and MCF-7Ras cell lines provide an isogenic model for the analysis of gene expression. Data show that H19 and E2F1 expressions were generally connected. E2F1 and H19 genes were weakly expressed in normal breast epithelial cells and up-regulated in BT20, T47D, and MCF-7 cells. Furthermore, both genes were activated in MCF-7Ras compared with MCF-7 cells. However, we noticed a dissociation in gene expression in MDA-MB-231, since E2F1 was highly expressed, whereas H19 was reduced to a low expression level. The relationship between E2F1 and H19 expression in breast cells led us to analyze gene expression in some cancer and normal breast tissues. We obtained an E2F1 overexpression in breast cancers (Fig. 6D). In the same way, the average H19 expression tended to be up-regulated in cancers even if to a varying extent, probably due to the cellular heterogeneity of breast tumors (Fig. 6E).
H19 Plays a Significant Role in the Cell Cycle Progression-E2F1 protein overexpression is sufficient to induce quiescent cells to enter the S-phase (56), and E2F1 is a potent transcriptional activator of E2F-responsive genes, such as Cdc6, cyclin E, and Cdc25, that are required for S-phase entry (34). In addition, the overexpression of some E2F target genes has been shown to promote S-phase entry in cells released from quiescence (57,58). Therefore, we considered that the H19 gene may play a role in S-phase entry in cells released from serum starvation. To test this hypothesis, we used H19-overexpressing cells. We previously characterized MDA-MB-231 cells, which contain a very low level of endogenous H19 RNA and were stably transfected with human H19 genomic sequences under the control of a heterologous promoter (26). These clones, which expressed a bona fide H19 RNA, were checked for H19 overexpression by real time RT-PCR. After normalization, S14-3 and S14-4 H19-transfected clones exhibited an H19 overexpression of 35-and 70-fold, respectively, when compared with mocktransfected cells (Neo) (Fig. 7, A and B). We investigated the effect of this H19 overexpression on the S-phase entry of these cells. Cells were starved and subsequently induced to reenter the cell cycle by serum addition. We measured the cell percentage in the S-phase by FACS analysis. At 18 -24 h, the S-phase percentage of H19-overexpressing cells (S14-3 and S14-4) was significantly higher than the percentage of the control cells (Neo) (Fig. 7C). In parallel, we monitored the H19 effect on cell growth. H19-transfected and Neo cells were plated (T0), serumstarved, and grown in a 10% FCS-containing medium for 48 h (T48). S14-3 and S14-4 clones displayed nearly twice as many cells as control cells (Fig. 7D). These results indicate that a H19 overexpression gives a new phenotype to cells, since it favors their S-phase entry after serum stimulation and confers an obvious advantage for cell proliferation.
Our studies have shown that H19 is maximally expressed during the S-phase and that the H19 overexpression accelerates S-phase entry in cells entering the cell cycle from quiescence ( Figs. 1 and 7, C and D). These results are consistent with the proposal that H19 plays a significant role in cell cycle progression. However, it remains to be determined whether H19 promotes cell cycle progression in asynchronously growing recombinant cells expressing ectopic H19. To address this point, we decided to specifically silence H19 expression in these cells by carrying out RNA interference assays. We tested two siRNA duplexes targeting H19 RNA and monitored the decrease in the H19 RNA level by real time PCR. As illustrated for the S14-4 clone (Fig. 8, A and B), transfection of these sequences reduced H19 RNA level by about 75% for siH19A FIG. 5. Cell cycle-regulated expression of the H19 promoter is dependent on E2F DNA-binding sites. MCF-7 cells were transfected with 1.2 g of luciferase reporter gene carrying either wild-type H19 promoter (WT) or H19 promoter mutated in the two E2F sites (DMut). 20 h post-transfection, cells were serum-starved for 24 h and subsequently stimulated with fresh medium containing 5% FCS. At the indicated times, the cells were harvested and assayed for luciferase and ␤-galactosidase activities in A and cell cycle distribution in B. The first lysate with H19 wild type promoter (0 h) was arbitrarily valued as 10 adjusted luciferase counts, whereas the others were calculated in comparison with that value. Transfection efficiencies were determined by pCMV-␤-galactosidase cotransfection. All values shown in A represent the mean values of three independent experiments. and 80% for siH19B compared with the control siGFP. We obtained the maximal inhibition (90%) by using a mixture of the two sequences, and consequently, this condition was applied for the following assays. To examine the specific effect of H19 expression on cell cycle progression, we performed growth assays with the asynchronously growing cells Neo, S14-3, and S14-4 and these latter two clones transfected with siRNA targeting either H19 or GFP as a negative control. As shown in Fig. 8C, H19 overexpression conferred a growth advantage on asynchronous cells (compare Neo and S14-4). In addition, siRNA that specifically targeted H19 sequences resulted in a significant decrease in cell growth compared with the control siRNA (compare S14-4siH19 and S14-4siGFP). We confirmed that H19 knock-down by siRNA was still efficient 72 h posttransfection. We obtained similar results with the other H19overexpressing clone (S14-3) (data not shown). Collectively, these results demonstrate that H19 is able to promote cell growth in breast cancer cells.
The inhibition of cell growth by the H19-specific siRNAs could result from a S-phase entry block or from an inhibition to progression through other phases of the cell cycle. In order to distinguish between these possibilities, we examined the cell cycle distribution of cells treated with H19-specific siRNA by FACS analysis in serum-starved cells. The siH19-treated cells were distributed over all phases of the cell cycle, with an accumulation in G 0 /G 1 -phase, since the percentage of cells is increased from 37 to 51% (Fig. 8D). In parallel, we performed growth assays using cell counting, and we obtained about a 20% decrease in the cell number after a 24-h starvation period, when cells were transfected with siH19 compared with the siGFP-transfected control cells (data not shown). Finally, we tested the effect of the endogenous H19 knock-down in MCF-7 cells to evaluate the physiological contribution of the gene to cell proliferation (Fig. 8E). RNA interference on H19 gene reduces the cell number by about 10% after 5 days of culture compared with the control siGFP sequences. Thus, our results demonstrate that the H19 mRNA-like noncoding RNA plays a significant function in favoring S-phase entry.

DISCUSSION
H19 is transcribed in an untranslated RNA molecule (8), which accumulates in the human placenta and several fetal tissues, and probably plays a pivotal role in embryogenesis and fetal growth and development (59). In addition, overexpression or reexpression of the gene occurring in numerous human cancers indicate that H19 is involved in oncogenesis (17).
Several lines of evidence presented in this study show that H19 is a bona fide E2F target gene and promotes cell cycle progression.
We report the cell cycle-dependent expression of the H19 gene in breast cancer cells. We demonstrated by chromatin immunoprecipitation assays that endogenous E2F1 was recruited to the H19 promoter in vivo and that E2F site mutation prevented complex formation in in vitro band shift experiments. These results conclusively show that a specific and direct E2F1 binding to E2F consensus sites can occur in the human H19 promoter. Transient transfection experiments in MCF-7 cells demonstrated that the H19 promoter is activated by E2F1 proteins through two FIG. 6. Connection between H19 and E2F1 expression patterns in breast cells. A, total RNA of normal breast epithelial cells (NBEC) and hormone-insensitive (BT20, MDA-MB-231, MCF-7Ras) or hormone-sensitive (T47D, MCF-7) cancer cell lines were isolated and reverse-transcribed. Real time PCR amplifications were performed using a SYBR Green mix. Relative E2F1 expression was calculated as indicated under "Materials and Methods," using the RPLP0 gene as reference. B, protein extracts of various cells were processed for Western blotting using an anti-E2F1 antibody and an anti-actin antibody as loading control. C, H19 expression was quantified by real time PCR and normalized to RPLP0 expression. D and E, total RNA were extracted from normal and cancer breast tissues described under "Materials and Methods." E2F1 RNA (D) and H19 RNA (E) expression levels were determined as mentioned above.
E2F recognition sequences present in the promoter. In functional assays, disruption by point mutation of E2F binding sites in the H19 promoter abolished the E2F1 ability to up-regulate gene transcription, confirming the potent role of both E2F recognition sequences. Thus, the E2F sites within the 5Ј regulatory region of the H19 gene are essential for gene induction. Moreover, we observed a H19 promoter repression by E2F1⌬, which may compete with endogenous wild-type proteins through its transdominant negative properties. In addition, E2F1 activity is required for endogenous H19 expression, as shown by RNA interference assays.
The H19 promoter was sensitive to another E2F member, E2F6, considered to be a transcriptional repressor, since it lacks both transactivation and pocket protein-binding domains. E2F6 was able to impede the endogenous E2F complex activity on the H19 promoter. This inhibition can be mediated either by exclusion of other E2F family members from the DNA, or through the recruitment of cellular factors that actively inhibit transcription (51,52). Furthermore, it is widely accepted that E2F family binding to pRb family proteins (pRb and related pocket proteins p107 and p130) is the primary regulatory mechanism for E2F family activity. For instance, E2F1 binds almost exclusively to the pRb member. The association between the two protein types forms inactive complexes, hence inhibiting E2F-dependent transcription (37,60). In agreement with these results, transfection in MCF-7 cells of a pRb expression vector abolished the H19 gene transactivation by E2F1. This finding brings supplementary proof of the H19 promoter control by E2F1.
The above results have clearly established the cell cycle-dependent expression of H19, and we further investigated the role of E2F sites present in the H19 promoter in this regulation. Indeed, mutations of these sites abolished the activation of H19 expression in S-phase during growth stimulation, clearly indicating the E2F1 control of H19 gene transcription during the cell cycle. Interestingly, the E2F1 gene, like H19, is broadly but not uniformly expressed during mouse embryogenesis (61) and is often deregulated in human cancers (42).
We therefore investigated the H19 and E2F1 expression patterns in various breast epithelial cells. The study revealed a correlation between H19 and E2F1 expression levels. Both genes were weakly expressed in normal breast cells and upregulated in breast cancer cells. We also used the isogenic model MCF-7/MCF-7Ras. MCF-7Ras cells, when stably transfected with the H-Ras oncogene exhibit high tumorigenic properties. This cell line lost its estrogen dependence, secreted diffusible growth factors that support its own tumor growth in vivo, and displayed an increased invasion capacity in in vitro assays (62,63). It is interesting to note that H19 and E2F1 expressions were activated in these cells when compared with the parental MCF-7 cells. This suggests that H19 and E2F1 expression levels match with the cancer cell aggressiveness. MDA-MB-231 cells are the only exception, where the expression of the two genes were dissociated. These cells exhibit a very low level of H19 transcripts but strongly expressed the E2F1 factor. Interestingly, we obtained an E2F1 overexpression in breast cancer tissues, and in the same way, the average H19 expression tended to be up-regulated in neoplasms. Thus, FIG. 7. H19 expression promotes cell proliferation. A, H19 overexpression in MDA-MB-231 clones. Total RNA of a mock-transfected (Neo) and the two H19-transfected clones (S14-3 and S14-4) were isolated and reverse-transcribed. Real time PCR amplifications were performed using a SYBR Green mix. The curves represent fluorescence versus cycle number of the target H19 gene and the reference RPLPO gene in Neo, S14-3, and S14-4 cDNAs. Cycle threshold determination was carried out at fluorescence level of 100. B, normalization of H19 expression. Relative H19 expression was calculated as indicated under "Materials and Methods" using the RPLP0 gene as reference. C, H19 promotes S-phase entry in MDA-MB-231 cells released from quiescence. MDA-MB-231 cells stably transfected with H19 were serum-starved for 48 h and subsequently stimulated with fresh medium containing 10% FCS. At the indicated times, the cells were harvested and assayed for cell cycle distribution by FACS. Neo (black bars) corresponds to a mock-transfected cell line, and both S14-3 (gray bars) and S14-4 (white bars) correspond to H19transfected cell lines. D, H19 accelerates MDA-MB-231 cell growth. Cells were plated (T0), serum-starved for 24 h, and grown in 10% FCS medium for 48 h (T48). All values shown represent the mean values of three independent experiments. it appears that gene expressions do not result in a single concomitant regulation. It is highly likely that other factors intervene in this complex regulation network. Consequently, we chose these cells to stably overexpress the H19 gene and to further investigate its function in cell growth.
Indeed, H19 overexpression in the MDA-MB-231 cell line accelerates the cell cycle progression, by increasing S-phase entry in asynchronously growing cells as well as in cells entering the cell cycle from quiescence. The inhibition of H19 expression by specific siRNA in these MDA clones suggests the direct contribution of the H19 gene to this cellular process and its involvement in a rate-limiting step in the G 1 /S-phase transition. Furthermore, knock-down of the endogenous H19 RNA in MCF-7 cells has a significant impact on cell proliferation. These data indicate that H19 is involved in cell cycle progression in physiological conditions. We can at least partly explain this function by our previous finding that thioredoxin is upregulated by H19 in MDA-MB-231 cells (13). Numerous studies argue for a role of thioredoxin in cell proliferation and transformation. This protein possesses a growth factor activity and has been reported to be overexpressed in a number of human primary cancers (64,65). Furthermore, thioredoxin transfected in breast cancer MCF-7 cells increased colony formation in soft agar and tumorigenesis in immunodeficient mice (66). The protein, added to minimal culture medium, stimulates the proliferation rate of these cells (67). Interestingly and in line with FIG. 8. H19 knock-down by RNA interference reduces cell proliferation. A, effect of siRNA sequences on H19 expression. S14-4 cells were transfected with 400 pmol of siH19A or siH19B alone or in combination for 48 h. Transfection of the siGFP sequence was used as a control condition. Total RNA were isolated and processed for real-time PCR. B, normalization of H19 expression. Relative H19 expression was calculated as indicated under "Materials and Methods" using the RPLP0 gene as reference. C, H19 inhibition reverse the growth phenotype of the H19-transfected cells. S14-4 cells were plated on day 1 (D1), transfected with siGFP (S14 -4siGFP) or the siRNA mix (S14 -4siH19) on day 2 (D2), and cultivated in 5% FCS medium until day 5 (D3-D5). In parallel, nontransfected S14-4 and Neo clones were used as positive controls. Statistical significance is indicated by asterisks as described under "Materials and Methods" and calculated for each day for S14 -4siH19 with regard to the corresponding control condition S14 -4siGFP. D, H19 promotes S-phase entry. S14-4 cells were transfected with the control siGFP or with the mix of siH19, serum-starved for 24 h, and submitted to a FACS analysis. The cell distribution in the different phases of the cell cycle is indicated as a percentage of total cells. E, RNAi on endogenous H19 in MCF-7 cells reduces cell proliferation. MCF-7 cells were plated on day 1 (D1), transfected with siGFP or the siRNA mix on day 2, and cultivated in 5% FCS medium until day 5 (D5), when cells were counted. Statistical significance is indicated by asterisks as described under "Materials and Methods" and calculated for D5 for siH19 with regard to the corresponding control condition siGFP. this result, Ayesh et al. (25) have shown a growth advantage for bladder carcinoma cells overexpressing H19 and cultured in serum-poor medium. This advantage was due in part to the inability of the H19-overexpressing cells to induce the cyclindependent kinase inhibitor p57 kip2 . The authors also revealed an up-regulation of cell cycle regulator genes, including genes involved in DNA synthesis, by H19. Therefore, considering all of these data and in agreement with the results presented here, H19 exerts its growth-promoting effect through its ability to modify the level of cell cycle molecules, and thioredoxin could be one of the mediators in this effect.
In conclusion, our present findings demonstrate that H19 promotes the G 1 -S transition in breast cancer cells through a link to the E2F1 factor. Fig. 9 is a comprehensive overview of our results. This work is the first documented report of the contribution of an mRNA-like noncoding RNA in cell cycle progression. Indeed, H19 can promote S-phase entry in mammalian cells under various conditions, and these observations provide the evidence that H19 expression is an active partner and not simply a consequence of cell proliferation. H19 overexpression confers an obvious advantage for cell proliferation and hence reinforces the aggressive phenotype of neoplastic cells. These properties explain the oncogenic function of H19 in breast cancer genesis.