An E2F1/MiR-17-92 Negative Feedback Loop mediates proliferation of Mouse Palatal Mesenchymal Cells

Normal cell cycle progression and proliferation of palatal mesenchymal cells are important for palatal development. As targets of miR-17-92, E2F transcription factors family has been suggested to induce the transcription of miR-17-92 in several cell types. In the present study, we sought to investigate whether this negative feedback loop exists in mouse PMCs and what the function of this negative feedback loop would be in palatal mesenchymal cells. Using GeneMANIA, we revealed that the most important function of experimentally verified targets of miR-17-92 is cell cycle regulation. E2F1 and E2F3, but not E2F2, were extensively expressed in mouse palate. Over-expression of E2F1 significantly increased the expression of all the members of miR-17-92. After increased by E2F1, miR-17 and miR-20a may negatively target E2F1, and thereby prevent the cells from excessive proliferation. We suggest that the negative feedback loop between E2F1 and miR-17-92 may contribute to palatal development by regulating the proliferation and cell cycle of palatal mesenchymal cells.

translation. In the past decades, the specific biological functions of miRNAs in specific cells or tissues have gained much attention. One of the best characterized polycistronic miRNAs is mir-17-92 7 . The miR-17-92 cluster is conserved among vertebrates, comprising six miRNAs: miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1 and miR-92a-1 8 . Originally found to be an onco-miRNA overexpressing in a variety of malignancies, miR-17-92 cluster has been demonstrated to function in a wide variety of settings, including normal development 8 . MiR-17~92−/− mice exhibit smaller size and die at birth, due to severe lung hypoplasia and cardiac defects 9 . In our previous study, miR-17-92 cluster has been found to continuously express in PMCs and palatal shelves in mouse embryo during E12-14, which is the critical period for palatal shelf elongation and elevation 10 . However, the mechanism by which miR-17-92 modulates palate development has been poorly understood.
An negative regulatory feedback loop between miR-17-92 cluster and E2F family has been described in Hela cells and neural stem cells 11,12 . This negative feedback loop between miR-17-92 cluster and E2F family is important for preventing an abnormal accumulation of E2F1-3 and may play a role in the regulation of cellular proliferation and apoptosis 11,12 . Because E2F family plays critical roles in cell cycle regulation, we hypothesized that there may be a negative feedback loop between miR-17-92 cluster and E2F family in PMCs, which may regulate the cell cycle of PMCs and palate development. In this paper, we show that E2F1 was increasingly expressed from E12 to E14 in PMCs. E2F1 induced miR-17-92 expression and proliferation of PMCs. In addition, miR-17 and miR-20a could directly target the 3′UTR of E2F1 in PMCs and imped G1/S transition of PMCs. We thus suggest that the negative feedback loop between miR-17-92 and E2F1 in PMCs plays important roles in regulating the cell cycle transition and proliferation of PMCs during the normal development of palate.

Results
Functional analysis of validated miR-17-92 target genes. To better understand the function of miR-17-92 cluster during palatal development, we identified the experimentally verified target genes of miR-17-92 using the miRTarBase 13 . A total of 91 genes were found to be miR-17-92 targets, among which, 41 of them are targeted by more than one components of miR-17-92. We further analyzed the function of these 91 genes and the functional association between them using GeneMANIA 14 . The most relevant function of these queried genes is the regulation of mitotic cell cycle. The functional associations between miR-17-92 target genes were shown in Fig. 1. As verified targets of miR-17-92, the E2F family, including E2F1, E2F2, and E2F3, plays central role in the functional network.
The sophisticated regulation of cell cycle is important during the embryonic development. However, how developmental cues controlling the cell cycle in different developmental situations remains to be addressed 15 . Because miR-17-92 and E2F family play pivotal roles in the cell cycle control, we hypothesized that the auto-regulatory feedback loop between mir-17-92 and E2F family may regulate palate development via cell cycle modulation.
To test whether E2F1 could regulate proliferation of PMCs, shRNA targeting E2F1 (shE2F1) was used to knock down E2F1. MTT assay and immunofluorescence staining of Ki-67 were applied to detect the cell proliferation. The knockdown of E2F1 significantly inhibited the proliferation of PMCs determined by MTT assay. ( Figure 2C) Moreover, the percentage of GFP/Ki67 double positive PMCs was significantly decreased by shE2F1 knockdown (Fig. 2D). These results suggest that E2F1 may regulate the proliferation of PMCs during the development of palate.
Positive regulation of miR-17-92 by E2F1. E2F1 regulates target gene expression usually via binding to the regulatory elements 11,12 . We sought to investigate whether E2F1 could regulate the expression of miR-17-92 in PMCs. PMCs were transfected with plasmids expressing E2F1 or GFP (control) respectively. Twenty-four hours after the transfection, RT-qPCR was used to measure the expression of miR-17-92 members. We found that over-expression of E2F1 significantly upregulated the expression of all the members of miR-17-92 (Fig. 3A).
Previous studies have shown that miR-17-92 had an E2F1consensus binding sequence on its regulatory elements 11,12 . Putative E2F1 consensus binding site were identified using "Sitescan" and "TFSEARCH". We used TTTSSCGC as an E2F1 binding sequence (where S = C or G) 16 , and looked for putative E2F1-binding sites by spanning −3 kb of genomic miR-17-92 cluster. Three putative binding sites were identified to match this canonical sequence. To confirm the direct regulation of E2F1 on miR-17-92, a ChIP assay was performed using PMCs. The miR-106a ORF region was used as negative control 11 . As shown in Fig. 3B, the PCR products of the second and third binding sites were detected when E2F1 was immunoprecipitated. These results validate that E2F1 could direct bind to the promoter of miR-17-92 and promote the transcription of miR-17-92in PMCs.  We next investigated whether mir-17-92 could directly down-regulate E2F1 in PMCs. Using the software 'TARGETSCAN' , "PICTAR" and 'MIRANDA' , we found that miR-17 and miR-20a have the same seed sequences and two predicted target sites in the same region of the 3′UTR of mouse E2F1 (Fig. 4C). However, we didn't find predicted target site of miR-18a/19a/19b/92a on E2F1. Thus we generated pMIR-Report luciferase vectors of E2F1 for miR-17/20a. The coding sequence of the firefly luciferase is followed by ~100 nucleotide synthetic DNA fragments encompassing the predicted miRNA binding site from predicted target gene 3′-UTR in either wild-type (WT) or seed mutant constructs. The 3′UTR recombinant construct of E2F1 was transfected into PMCs along with miR-17/20a mimics or scrambled miRNAs. The luciferase activity in PMCs transfected with E2F1 WT constructs plus miR-17/20a mimics was significantly lower than that in PMCs transfected with either E2F1 WT or mutant constructs alone, and the scrambled miRNA did not affect the luciferase activity in either WT or mutant constructs transfected PMCs (Fig. 4D). This result indicates that E2F1 is a direct target of miR-17 and miR-20a in PMCs.

MiR-17-92 negatively regulates E2F1-induced cell cycle transition of PMCs. To investigate
whether miR-17-92 could regulate cell cycle of PMCs through targeting E2F1, PMCs were transfected with miR-17-92 mimics, miR-17-92 inhibitors, and scrambled miRNAs respectively. The proliferation rate was measured up to 72 h after transfection using MTT assay. We found that PMCs transfected with miR-17-92 mimics had a significantly lower cell proliferation rate than that transfected with scrambled miRNAs. Cells transfected with miR-17-92 inhibitors had a significantly higher proliferation rate compared with those transfected with scrambled miRNAs (Fig. 5A).
Flow cytometry analysis was used to examine the cell cycle of PMC with PI staining (Fig. 5B). PMCs transfected with miR-17-92 mimics had a higher rate of G0/G1 phase (Fig. 5C) while lower rate of S phase compared with those transfected with scramble miRNA (Fig. 5D). To investigate whether this effect was dependent on miR-17 and miR-20a, the same cells were co-transfected with miRNA inhibitors against miR-17 and miR-20a respectively. The inhibition of miR-17 and miR-20a in cells transfected with miR-17-92 mimics decreased the proportion of cells in G0/G1 phase (Fig. 5C) while increasing the percent of cells in S phase (Fig. 5D). These results suggest that the effect of miR-17-92 on cell cycle control is dependent on miR-17 and miR-20a which directly target E2F1 in PMCs.

Discussion
MiR-17-92 has been found to affect palatal development 17,18 . In the present study, using GeneMANIA, we revealed that the most important function of experimentally verified targets of miR-17-92 is cell cycle regulation.
The functional network provides a framework in which large datasets are analyzed with an unbiased view and their functions are better understood 19 . Also, the network framework is a powerful concept and tool for revealing molecular mechanisms and predictive biomarkers 20   As a target of miR-17-92, E2F1 has been suggested to induce the transcription of miR-17-92 in several cell types. We sought to investigate whether this negative feedback loop exists in mouse PMCs and what the function of this negative feedback loop would be in PMCs. Our results showed that E2F1 and E2F3, but not E2F2, were extensively expressed in mouse palate. E2F1 could induce the proliferation of PMCs and the expression of miR-17-92. After increase, miR-17 and miR-20a may negatively target E2F1, and thereby prevent the cells from excessive proliferation. We suggest that the negative feedback loop between E2F1 and miR-17-92 may contribute to palatal development by regulating the proliferation and cell cycle of PMCs (Fig. 6).
In the process of palate development, the secondary palate primordium extends from the oral surface of the maxillary processes to form palatal shelves. Before palatal fusion, palate development includes palatal shelf elongation and elevation. The normal process of these two steps demands normal cell cycle progression of PMCs. Here we found that E2F1 expressed in palate during E12 to E14. Moreover, E2F1 induces proliferation of PMCs ex-vivo. Our results suggest that E2F1 in PMCs may contribute to the continuous growth of palate during E12 to E14. E2F2 and E2F3 have been found to express in palatal shelves. In this study, however, the E2F2 expression level was extremely low. The reason might be that E2F1 and E2F2 function at different stage of palate development.
It is well known that E2Fs transcription factors are activated by hypo-phosphorylation and gradual inactivation of Rb by cyclins and CDKs. This process triggers a positive feedback loop. More cyclin/CDK would result in hyper-phosphorylation and complete inactivation of Rb, which activates more E2Fs. This positive feedback mechanism commits a cell to pass the G1/S transition and get into S phase 22 . However, once the palatal shelves are fused, the high proliferation nature of PMCs would be terminated, and the anterior portion undergoes ossification to form the hard palate 23 . Therefore, the activity of E2F1 is needed to be balanced by a negative signal. Mir-17-92, directly induced by E2F1 12,24 , may act as the negative signal by directly target E2F1. This negative regulation between E2F1 and miR-27-92 may precisely maintain the cell cycle transition and proliferation of PMCs during the development of palate. Abnormal regulation of this negative feedback may result in developmental disorder of palate.

Materials and Methods
Cell culture. All the animals involved in our study were approved by the Animal Care and Use Committee of Sichuan Cancer Hospital. All of the experimental procedures followed by Guide for the Care and Use of Laboratory Animals: Eighth Edition (NIH, Bethesda, MD, USA). We dissected palatal shelves from C57BL/6J mouse embryos on E13.5, and separated the mesenchyme from the epithelia. To dissociate individual cells, the palatal mesenchyme was incubated in PBS with 0.25% trypsin and 0.02% EDTA for 15 min. Primary culture of PMCs was initiated by seeding 5 or 2 ml DMEM/F12 into 25 ml flask or flat plate.
Quantitative real-time RT-PCR. Reverse transcription and qPCR of the components of miR-17-92 cluster was performed using Taqman Small RNA Assay (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's instructions. Briefly, 10 ng total RNA was reverse transcribed with miRNA specific primers in 15 ml reaction volumes. Reverse transcription reactions were diluted and amplified in triplicates by TaqMan qPCR on a 7300 Real Time PCR System (Applied Biosystems). Quantification was performed using the ΔΔCt method. The RT-PCR fold changes were normalized to snoRNA135 (Applied Biosystems).
For E2F1/2/3, quantitative real-time RT-PCR was carried out as described previously 25 . Briefly, total RNA was isolated with TRIzol reagent (Invitrogen, Carlsbad, CA, USA), and reverse-transcribed using a RevertAid First-Strand cDNA Synthesis Kit (Fermentas). PCR amplification of the cDNA template was done using Thunderbird SYBR qPCR mix (Toyobo, Osaka, Japan) on ABI PRISM 7300 sequence detection system (Applied Western blot. Western blots of E2F1, E2F2 and E2F3 were carried out as described previously 26 . Briefly, total proteins were isolated from the PMCs. Thirty-microgram proteins from each sample were separated on SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore, Darmstadt, Germany). Membranes were incubated with anti-E2F1 (Santa Cruz, Dallas, TX, USA), anti-E2F2 (Santa Cruz) and anti-E2F3 (Abcam, Cambridge, MA, USA) antibodies.
MTT assay for cell proliferation. The cell proliferation was quantified by the colorimetric MTT assay as previous described 27 . In brief, cells were incubated with MTT for 4 h. Then supernatant was removed and DMSO was added. Optical densities at 490 nm were measured using culture.
Flow cytometry. PMCs transfected with miR-17-92 mimics/inhibitors/scramble were synchronized at the G0/G1 phase of the cell cycle by serum starvation for 24/48 h with 0.5% FBS 29 . Then cells were harvested and stained with propidium iodide (PI) for cell cycle analysis using Click-iT EDU flow cytometry assay kit (Invitrogen, CA, USA) on a Cytomics FC 500 MPL Flow Cytometer (Beckman Coulter, Brea, CA, USA) with RXP software (Beckman Coulter). Data were analyzed using MODFIT LT 4.1 software.