Transcriptional Repression of Cdc25B by IER5 Inhibits the Proliferation of Leukemic Progenitor Cells through NF-YB and p300 in Acute Myeloid Leukemia

The immediately-early response gene 5 (IER5) has been reported to be induced by γ-ray irradiation and to play a role in the induction of cell death caused by radiation. We previously identified IER5 as one of the 2,3,4-tribromo-3-methyl-1-phenylphospholane 1-oxide (TMPP)-induced transcriptional responses in AML cells, using microarrays that encompassed the entire human genome. However, the biochemical pathway and mechanisms of IER5 function in regulation of the cell cycle remain unclear. In this study, we investigated the involvement of IER5 in the cell cycle and in cell proliferation of acute myeloid leukemia (AML) cells. We found that the over-expression of IER5 in AML cell lines and in AML-derived ALDHhi (High Aldehyde Dehydrogenase activity)/CD34+ cells inhibited their proliferation compared to control cells, through induction of G2/M cell cycle arrest and a decrease in Cdc25B expression. Moreover, the over-expression of IER5 reduced colony formation of AML-derived ALDHhi/CD34+ cells due to a decrease in Cdc25B expression. In addition, over-expression of Cdc25B restored TMPP inhibitory effects on colony formation in IER5-suppressed AML-derived ALDHhi/CD34+ cells. Furthermore, the IER5 reduced Cdc25B mRNA expression through direct binding to Cdc25B promoter and mediated its transcriptional attenuation through NF-YB and p300 transcriptinal factors. In summary, we found that transcriptional repression mediated by IER5 regulates Cdc25B expression levels via the release of NF-YB and p300 in AML-derived ALDHhi/CD34+ cells, resulting in inhibition of AML progenitor cell proliferation through modulation of cell cycle. Thus, the induction of IER5 expression represents an attractive target for AML therapy.


Introduction
Acute myeloid leukemia (AML) is characterized by the excess production of leukemic blasts arrested at various stages of granulocytic and monocytic differentiation. To effectively cure a patient with AML, this proliferation of leukemic cells must be halted. Given that chemotherapy rarely eradicates the leukemic clones, efforts are now being made to find innovative new therapies which inhibit the proliferation of AML cells. However, the effect of cell cycle progression and apoptosis resistance on the pathogenesis of AML remains to be defined. Against these backgrounds, we have synthesized new bioactive agents and then investigated these antileukemic effects. We previously reported that the phospha sugar derivative, 2,3,4-tribromo-3-methyl-1-phenylphospholane 1-oxide (TMPP), was synthesized in the reaction of 3-methyl-1-phenyl-2phospholene 1-oxide with bromine, and we investigated the potential of TMPP as an anti-leukemic agent using AML-derived ALDH hi cells [1]. This agent induced a G2/M cell cycle block through a reduction in cell cycle progression signals (FOXM1, KIS, Cdc25B, Cyclin D1, Cyclin A, and Aurora-B), resulting in inhibition of leukemia cell proliferation [1]. We also observed that down-regulation of FOXM1 inhibited proliferation, and demonstrated that TMPP suppressed FOXM1 expression, and that this FOXM1 repression reduced Cyclin B1 and Cdc25B mRNA expression, resulting in inhibition of the proliferation of AMLderived ALDH hi cells [2]. Thus, we demonstrated that TMPPmediated FOXM1 repression induced G2/M cell cycle arrest through a reduction in Cyclin B1 and Cdc25B expression. However, TMPP and FOXM1 regulate many mitotic regulators in AML cells. It is unclear how TMPP predominantly induces G2/M cell cycle arrest rather than G1 cell cycle arrest in AML cells.
To identify TMPP-induced transcriptional responses in AML cells, TMPP-induced transcriptional alterations were investigated using microarrays that encompassed the entire human genome.
About 180 genes, which belong to functional categories such as the DNA damage response, regulation of cell cycle and cell proliferation, and signaling pathways, responded to TMPP treatment at the transcriptional level in AML cells. Of these genes, the immediate-early response gene 5 (IER5) was identified as a key regulator of the G2/M cell cycle transition.
The immediate-early genes (IER), which are rapidly induced by growth factors or other various stimuli, encompass a variety of different protein families (Fos and Jun family of transcriptional regulators; Myc; zinc-finger proteins; secreted cytokines; cytoplasmic proteins, and integral membrane proteins) [3]. Activation of IER is an important initial step in the regulation of cellular and genomic responses to external stimuli. Approximately 100 IER genes have been described to date, and are subdivided into two classes (fast-kinetics and slow-kinetics) based on their activation kinetics [4]. The fast-kinetics IER genes (e.g., c-Fos) contain serum response elements (SRE), which are required for transcriptional induction. In contrast, the slow-kinetics IER genes, which lack SRE, display a relatively slower induction and longer persistence profile following stimulation compared with the fast-kinetics IER genes [5]. The IER5 gene, which has been identified as a member of the IER gene family, belongs to the slow-kinetics IER genes, and is rapidly induced by stimulation with serum or with the growth factors FGF or PDGF [6]. It has been also reported that IER5 mRNA is induced in the cerebral cortex of rats during waking and sleep deprivation [7], or in the brains of mouse embryos exposed to teratogenic valpronic acid (VPA) [8]. The IRE5 mRNA was induced within 30 min after serum-exposure and at least 180 min after the serum-stimulation, but its expression was not inhibited by cycloheximide [6]. IER5 is also upregulated by ionizing radiation at doses ranging from 0.02 to 10 Gy in lymphoblastoid AHH-1 cells [9,10]. Moreover, it has been reported that suppression of IER5 increased HeLa cell proliferation, mitigated the inhibition of proliferation imposed by irradiation, and potentiated radiationinduced arrest at the G2/M transition [11]. These results demonstrated that IER5 expression plays an important role in radiation-mediated cell death and cell cycle checkpoints.
It has been reported that inhibition of cell proliferation in AML cells is associated with a decrease in the expression of the Cdc25B phosphatase [12], and that this phosphatase participates in G2/M checkpoint recovery and its expression is upregulated in acute myeloid leukemia cells [13]. Therefore, depletion of Cdc25B might be expected to strongly induce G2/M cell cycle arrest in AML cells. As previously reported, siRNA-mediated depletion of FOXM1 expression, or TMPP treatment, induced G2/M cell cycle arrest and inhibited AML cell proliferation through a decrease in protein expression of mitotic regulators such as Cdc25B in AML cells [1,2]. However, it is unclear how Cdc25B expression is regulated.
In this study, we investigated the function of IER5 in leukemia cell proliferation. We found that IER5 expression inhibited the proliferation of both leukemia cell lines and of leukemic blast cells derived from AML through the transcriptional repression of Cdc25B.

Results
The expression of IER5 mRNA in acute myeloid leukemia cells We first determined the relative expression of IER5 mRNA in the leukemia cell lines KG-1, Kasumi-1, U937 and YRK2. As shown in Fig. 1, IER5 mRNA was constitutively expressed in these AML cell lines. Interestingly, we found that the mRNA expression of IER5 increased in these AML cell lines compared to untreated cells, when treated with TMPP (5 and 10 mM) for 24 h. However, Ara-C (1 mM) did not affect IER5 mRNA expression in these AML cell lines. Furthermore, we quantified the level of IER5 mRNA in the AML cell lines using quantitative RT-PCR. This analysis indicated that the expression of IER5 mRNA in the TMPP-treated leukemia cells was increased 1.78 to 2.29-fold relative to its expression in untreated cells.
We next determined the relative expression levels of IER5 mRNA and protein in U937 cells. As shown in Fig. 1B, both IER5 mRNA and protein in U937 cells were overexpressed following transfection with IER5 cDNA or treatment with TMPP compared to untransfected and untreated U937 cells, respectively. In contrast, transfection with IER5 shRNA-#1 or -#2 decreased the expression of both IER5 mRNA and protein in U937 cells compared to untransfected U937 cells. Moreover, the mRNA and protein expression of IER5 in other AML cell lines (KG-1, Kasumi-1 and YRK2) was similarly increased by IER5 cDNA transfection and suppressed by IER5 shRNA transfection (data not shown). Thus, IER5 mRNA and protein were constitutively expressed in all four AML cell lines, although their expression in AML cell lines were at lower levels compared to normal ALDH hi / CD34 + cells (data not shown). Treatment with TMPP increased both the mRNA and the protein level of IER5 in AML cells.

Effects of IER5 expression on AML cell proliferation
Since IER5 expression was induced by TMPP in AML cells, we next examined the functional importance of IER5 expression. For this purpose, we transfected U937 cells with IER5 cDNA, and assessed the effect of IER5 over-expression on AML cell proliferation over 72 h of culture, starting from day 3 posttransfection. Cell proliferation was measured by cell counting using a hemocytometer ( Fig. 2A, upper panel). When U937 cells were transfected with IER5 cDNA, cell proliferation was inhibited compared to non-transfected control cells. Treatment with TMPP (5 mM) also significantly inhibited the proliferation of U937 cells. However, the viability of AML cells transfected with IER5 cDNA or treated with TMPP was almost same as that of the untreated cells ( Fig. 2A, bottom panel). In addition, the expression of IER5 mRNA was significantly increased by TMPP treatment compared to untreated cells (Fig. 2B, upper panel). We further examined the effect of IER5 induction on AML cell proliferation by flow cytometric analysis of the effect of IER5 on cell cycle distribution. As shown in Fig. 2C and 2D, analysis of U937 cells on day 3 post cDNA-transfection or post-TMPP treatment indicated an increase in the percentage of cells in S and G2/M, but not in sub-G1, compared to control DNA-transfected and untreated cells. No increase in the polyploid (.4N) population of these cells was observed. On the other hand, when U937 cells were transfected with IER5 shRNA-#1 or -#2, each cell cycle distribution was not affected compared to scrambled shRNA-transfected cells. In contrast, the population of G2/M was significantly reduced following TMPP treatment. Moreover, we determined the effect of IER5 over-expression and TMPP treatment on markers of apoptosis. We analyzed the effects on loss of mitochondrial membrane potential, which was determined by flow cytometric analysis of DiOC6 uptake. Loss of mitochondrial membrane potential is known to occur in apoptotic cells and precedes the activation of caspases. DiOC6 fluorescence of U937 cells was not reduced 3 days after transfection with IER5 cDNA or TMPP treatment, and DiOC6 fluorescence of IER5 cDNA-transfected or TMPP treated cells compared to control cells was observed in the almost same levels (Fig. 2E). The IER5 over-expression and TMPP treatment similarly induced G2/M cell cycle arrest, but not apoptosis in other AML cell lines (KG-1, Kasumi-1 and YRK2) (data not shown). These results showed that induction of IER5 induced a block in the G2/M phase in AML cells, but did not induce apoptosis, and therefore demonstrated that induction of IER5 expression strongly inhibited AML cell proliferation.

Effects of IER5 knock-down on AML cell proliferation
To further examine the functional importance of IER5 expression in AML cells, we next examined the effect of IER5 knock-down on AML cell proliferation. U937 cells were transfected with IER5 shRNA-#1 or -#2, and the effect of IER5 knock-down on AML cell proliferation was assessed over 72 h of culture, starting from day 3 day post-transfection ( Fig. 3). Cell proliferation was measured by counting the cells using a hemocytometer (upper panels). When U937 cells were transfected with IER5 shRNA-#1 or -#2 ( Fig. 3A and 3B, respectively), cell proliferation was slightly increased compared to both untreated  . Over-expression of IER5 inhibited AML cell proliferation. A. The cell proliferation of U937 cells, transfected with IER5 cDNA or treated with TMPP, was measured by counting cells using a hemocytometer (upper panel). Cell counting was started 3 days after transfection, and was performed every 24 h for 3 days. Data are shown as means 6 S.D. of triplicate cultures and are representative of three independent experiments. *P,0.01 compared with untransfected control cells. Cell viability of the IER5 over-expressing U937 cells was assessed by counting of viable cells using trypan blue staining at 72 h, starting 3 days after DNA transfection (bottom panel). B. QRT-PCR analysis of IER5 mRNA expression in untreated cells, IER5 cDNA-transfected cells, and TMPP-treated U937 cells. QRT-PCR was started 3 days after transfection, and was performed every 24 h for 3 days. Data are shown as means 6 S.D. of triplicate cultures and are representative of three independent experiments. The levels of the quantified RT-PCR products were normalized to GAPDH expression in the same sample and were then expressed relative to the mRNA level of a normal control, which was assigned a value of 1. *P,0.01 compared with untransfected control cells. The protein expression of IER5 in cells was analyzed after 3 days of culture (bottom panels). Blotting of Actin was used as a loading control. C. The cell cycle distribution of U937 cells that were transfected with control DNA, IER5 cDNA, scrambled shRNA, IER5 shRNA #1, IER5 shRNA #1, or were treated with TMPP, was analyzed using flow cytometric analysis. The transfected or TMPP-treated U937 cells were harvested after 3 days. The fraction of cells in the G1, S and G2/M stage of the cell cycle was determined. Data are shown as means 6 S.D. of triplicate cultures. The IER5 mRNA expression of the cells is shown at bottom and was assessed using RT-PCR. The RT-PCR results are representative of three independent experiments. Data are shown as means 6 S.D. of triplicate cultures. *P,0.01 compared with control cells. D and E. Cell cycle analysis (D) and changes of mitochondrial membrane potential (DYm) (E) in IER5 overexpressed or TMPP treated AML cells. U937 cells were transfected with IER5 cDNA or treated with TMPP (5 mM). Mitochondrial membrane potential was determined 3 days after transfection or TMPP treatment by staining of the cells with DiOC6 followed by flow cytometric analysis. The FACS results are representative of three independent experiments. NC; Negative control. doi:10.1371/journal.pone.0028011.g002 cells and to scrambled shRNA-transfected cells. In contrast, cell proliferation was significantly reduced following TMPP treatment. However, TMPP-mediated inhibition of the proliferation of U937 cells was reduced by knock-down of IER5 via transfection of IER5 shRNA-#1 or -#2. Moreover, TMPP-mediated IER5 induction was decreased in U937 cells transfected with IER5 shRNA-#1 or -#2 (bottom panels) compared to non-transfected cells. Knockdown of IER5 in other AML cell lines (KG-1, Kasumi-1 and YRK2), similarly increased cell proliferation (data not shown). These results show that TMPP inhibits AML cell proliferation through IER5 induction.

IER5 induced cell cycle arrest in AML cells through direct binding to Cdc25B promoter and decrease of NF-YB binding
We next determined if the inhibitory effect of IER5 induction on cell proliferation was mediated by modulation of the expression of cell cycle regulators such as Cdc25B, CHK1, WEE1, Cyclin B1, Survivin, KIS, or Aurora B kinase, during the G2/M transition. As shown in Fig. 4A, induction of the IER5 protein in IER5 cDNA-transfected or in TMPP-treated U937 cells was accompanied by a significant decrease in the level of Cdc25B protein compared to untransfected cells. IER5 over-expression did not affect CHK1, WEE1, Cyclin B1, Survivin or Aurora B kinase expression. In contrast, TMPP treatment reduced the expression of Cyclin B1, Survivin, and Aurora B kinase. TMPP treatment did not affect CHK1 and WEE1 expression. Induction of IER5 similarly reduced Cdc25B expression in other AML cell lines (KG-1, Kasumi-1 and YRK2) (data not shown). These results indicate that, during TMPP-mediated G2/M cell cycle arrest, TMPP reduced Cdc25B expression through IER5 induction and that IER5 is essential for the repression of Cdc25B protein expression in AML cells. Moreover, IER5 only regulated Cdc25B, and did not affect other cell cycle regulators of the G2/M transition.
To confirm binding of IER5 to the Cdc25B promoter and more accurately define its binding site, we performed ChIP analysis in AML cells using two pairs of oligonucleotides within upstream from the transcription start site for Cdc25B (Fig. 4B). ChIP analysis showed that IER5 is highly enriched within a region of the promoter of Cdc25B, indicating that the Cdc25B gene is a target for IER5 (Fig. 4C). Moreover, to determine whether the overexpression of IER5 corresponded to the binding sites of Sp1, Sp3, and NF-YB within the Cdc25B promoter, ChIP analysis was performed using antibodies corresponding to Sp1, Sp3, and NF-YB, which has been reported to be related to p53-mediated negative regulation of Cdc25B [14]. As shown in Fig. 4D, the reduced binding of NF-YB on the Cdc25b promoter was observed in IER5 over-expressed U937 cells compared to untreated and control DNA transfected cells. On the other hand, the levels of binding Sp1 and Sp3 on the Cdc25B promoter was not affected in untreated, control DNA-, or IER5-transfected cells. In addition, we investigated whether the reduction of NF-YB binding to Cdc25B promoter interferes with the recruitment of coactivator p300, which is anti-histone acetyltransferase as a coactivator known interact with NF-Y [15]. When U937 cell was untreated or transfected with control DNA, p300 was bound on the Cdc25B promoter. When IER5 over-expressed in U937 cells, a significant release of p300 was observed (Fig. 4E). On the other hand, the DNA methyltransferase DNMT1, which acts as a loading platform for repressor complexes such as the p53-Sp1 complexes [16] was not recruited on the Cdc25B promoter in IER5 over-expressed U937 cells. These ChIP experiments show that the binding of IER5 on the Cdc25B promoter induced the release of the coactivator p300, but not the recruitment of the coactivator DNMT1 via Sp1 interaction, causing the down-regulation of Cdc25B expression.

Induction of IER5 inhibited colony formation of AML derived-ALDH hi /CD34 + cells
We next examined IER5 mRNA expression in clinical specimens from AML patients. Hematopoietic progenitor cells from bone marrow were obtained using flow cytometry based on ALDH activity, using the substrate Aldefluor as previously reported. The CD34 + progenitor cells in the ALDH hi population (ALDH hi /CD34 + cells) were then sorted using FACS. We isolated ALDH hi /CD34 + cells from bone marrow cells derived from 2 healthy volunteers (#1 and #2) and 2 AML patients (M1 and M2) ( Fig. 5A and B). The percentage of bone marrow cells derived from healthy volunteers #1 and #2 that were ALDH low was 98.6% and 97.1%, and that of ALDH hi cells was 0.37% and 0.52%, respectively. The percentage of the ALDH hi cells that was CD34 + was 84.2% and 81.9%, and that was CD34was 14.7% and 17.5%, respectively. On the other hand, in bone marrow cells derived from AML patients M1 and M2, the percentage of ALDH low cells was 85.4% and 81.1%, respectively, and that of ALDH hi cells was 7.4% and 6.7%, respectively. The percentage of the ALDH hi cells that was CD34 + was 87.1% and 80.6%, and that was CD34was 11.4% and 19.3%, respectively.
To confirm that IER5 also modulated the proliferation of AML progenitor cells from patients, we analyzed the effect of IER5 expression on colony formation of ALDH hi /CD34 + cells from normal healthy volunteer #1 and from an AML patient (M1) prior to treatment (Fig. 5C). Transfection with IER5 cDNA or treatment with TMPP induced dramatic decreases in the number of colonies formed by AML-derived ALDH hi /CD34 + cells compared to the number in untreated cells ( Fig. 5C; left panels). In untransfected AML (M1) cells, the mean number of colonies was 112 (range, 103-119). Following transfection with IER5 cDNA and treatment with TMPP the mean number of colonies was 27 (range, [21][22][23][24][25][26][27][28][29][30][31][32][33][34][35] and 16 (range, 12-27), respectively. In contrast, transfection with IER5 cDNA only moderately decreased the number of colonies formed by normal ALDH hi /CD34 + cells, compared to untransfected normal ALDH hi /CD34 + cells. Thus, the mean number of colonies in untransfected cells was 46 (range, 39-52) but that in cells transfected with IER5 cDNA was 22 (range, [18][19][20][21][22][23][24][25][26][27][28][29]. These results demonstrate that induction of IER5 expression predominantly inhibited the proliferation of AML-derived ALDH hi / CD34 + cells. In addition, to confirm that IER5-mediatedinhibition of colony formation of AML-derived ALDH hi /CD34 + cells from patients was mediated by an effect on transcription of the IER5 target gene, we assayed the mRNA expression of the IER5 target gene Cdc25B in AML-derived ALDH hi /CD34 + cells. As shown in Fig. 5C (right panels), induction of IER5 expression significantly decreased the expression of Cdc25B mRNA in AMLderived ALDH hi /CD34 + cells, and also decreased the expression of Cdc25B mRNA in normal ALDH hi /CD34 + cells. Moreover, in AML (M1) ALDH hi /CD34 + cells, IER5 over-expression or TMPP treatment indicated an increase in the percentage of cells in S and G2/M compared to control cells (Fig 5D). No increase in the sub-G1 population of these cells was observed. Interestingly, we also detected an increase (3.1 to 4.2%) in the percentage of polyploid cells (.4N) in TMPP treated cells compared to control cells. In addition, as shown in Fig. 5E, DiOC6 fluorescence of AML (M1) ALDH hi /CD34 + cells was not reduced 3 days after transfection with IER5 cDNA or TMPP treatment, and DiOC6 fluorescence of IER5 cDNA-transfected or TMPP treated cells compared to control cells was observed in the almost same levels. Similar results were obtained using other AML ALDH hi /CD34 + cells. (data not shown). These findings suggest that the induction of IER5 expression strongly inhibited colony formation and G2/M cell cycle arrest in AML-derived ALDH hi /CD34 + cells compared to normal ALDH hi /CD34 + cells through induction of Cdc25B expression.
Colony forming activity of ALDH hi /CD34 + cell populations from AML patients following the induction or depletion of IER5 expression We examined the effect of induction or knock-down of IER5 on the colony formzing activity of ALDH hi /CD34 + hematopoietic progenitor cells from an AML patient (M1) (Fig. 6A and C). When cells were left untreated, the mean number of colonies was 117 (range, 102-126). When the cells were transfected with IER5 cDNA and/or treated with TMPP, the mean numbers of colonies were 7 (range, 3-11), 28 (range, 20-37), and 22 (range, 14-26), respectively. Thus colony numbers were dramatically reduced when ALDH hi /CD34 + cells were cultured with TMPP and/or were transfected with IER5 cDNA. In ALDH hi /CD34 + cells transfected with IER5 shRNA-#1 or -#2, TMPP moderately reduced colony numbers compared to untransfected cells. Moreover, in each colony derived from ALDH hi /CD34 + cells that was treated with TMPP or transfected with IER5 cDNA, the expression of IER5 mRNA was increased and the expression of Cdc25B mRNA was significantly decreased compared to untreated cells (Fig. 6B). These results demonstrate that during the proliferation of progenitor cells derived from AML patients, IER5 induced a reduction in Cdc25B expression.
To further assess the effects of IER5-mediated Cdc25B reduction on colony formation of ALDH hi /CD34 + cells from an AML patient (M1), we investigated whether transfection of Cdc25B DNA could rescue the colony forming activity of the AML-derived ALDH hi /CD34 + cells transfected with IER5 DNA or treated with TMPP. ALDH hi /CD34 + cells were transfected with IER5 cDNA or were treated with TMPP. As shown in Fig. 6D and F, either TMPP treatment or IER5 over-expression inhibited colony formation of AML-derived ALDH hi /CD34 + cells. In contrast, over-expression of Cdc25B in ALDH hi /CD34 + cells increased the number of colonies compared to untransfected cells. Moreover, the expression level of IER5 mRNA was not affected in any colony derived from ALDH hi /CD34 + cells that was transfected with Cdc25B DNA (Fig. 6E). Similar results were obtained using AML (M2) ALDH hi /CD34 + cells. These findings suggest that the transcriptional repression mediated by IER5 regulated Cdc25B expression levels, and, subsequently, decreased the expression of Cdc25B, resulting in inhibition of colony formation.
The reduction of NF-YB binding to the Cdc25B promoter by IER5 causes the decrease of Cdc25B expression Finally, to confirm binding of IER5 to the Cdc25B promoter in AML-derived ALDH hi /CD34 + cells, we also performed ChIP analysis in ALDH hi /CD34 + cells purified from one AML patient (AML: M1) (Fig. 7). ChIP analysis showed that IER5 is also highly enriched within a region of the promoter of Cdc25B (Fig. 7, left   panel). Moreover, to determine whether the over-expression of IER5 corresponded to the binding site of NF-YB within the Cdc25B promoter, ChIP analysis was performed using the anti-NF-YB antibody. The result shows that the reduced binding of NF-YB on the Cdc25b promoter was observed in IER5 over-expressed AML-derived ALDH hi /CD34 + cells compared to untreated and control DNA transfected cells (Fig. 7, middle panel). In addition, a significant release of p300 was also observed in IER5 overexpressed AML-derived ALDH hi /CD34 + cells, (Fig. 7, right  panel). These results show that the binding of IER5 on the Cdc25B promoter induced the release of the coactivator p300, causing the down-regulation of Cdc25B expression in AMLderived ALDH hi /CD34 + cells.

Discussion
We investigated the role of IER5 in inhibition of the growth of AML cells induced by treatment with TMPP. We demonstrated, for the first time, that TMPP induces the expression of IER5 and inhibits AML cell proliferation via IER5-mediated transcriptional repression of Cdc25B. Combined with the results of our previous study that TMPP-mediated FOXM1 repression induced G2/M cell cycle arrest, our results provide in vitro evidence to support a role for FOXM1 as an oncogene in AML cells, and to support the fact that its down-regulation inhibits the proliferation of established AML cell lines and primary AML cells.
IER5, which encodes a 308-amino acid member of the immediate-early class of signal transduction proteins, is induced by growth factors and mediates the cellular response to mitogenic signals [6]. IER5 also functions as a transcriptional regulator by binding to DNA or by mediating nuclear protein-protein interactions [6]. Moreover, it has been reported that ionizing radiation induces IER5 mRNA expression in a dose-and time-dependent manner, and that suppression of IER5 increases the proliferation of HeLa cells and mitigates the inhibition of proliferation induced by exposure to radiation, thereby increasing the resistance of HeLa cells to radiation. In contrast, increased expression of IER5 reportedly disturbed the cell cycle checkpoint and sensitized cells to radiation [11]. These previous results have shown that IER5 expression plays a role in the induction of cell death that is caused by radiation. However, the biochemical pathway and mechanisms of IER5 function in regulation of the cell cycle remain unclear. In this study, we investigated the involvement of IER5 in the cell cycle and in proliferation of AML cells and AML-derived leukemic progenitor cells. We found that treatment with TMPP induced the expression of IER5 mRNA and that over-expression of IER5 in AML cells inhibited their proliferation. Moreover, suppression of IER5 expression in AML cells using shRNA increased the proliferation of AML cells, and mitigated TMPP-mediated inhibition of proliferation compared to untransfected AML cells. We also found that the both IER5 over-expression and induction of IER5 expression by TMPP treatment induced an increase in the population of AML cells and primary AML blastic cells in the . IER5 induced G2/M cell cycle arrest through direct binding to Cdc25B promoter in AML cells. A. U937 cells were transfected with control DNA, IER5 cDNA, or were treated with TMPP (5 mM or 10 mM). The cells were harvested after 3 days, and the expression of the indicated cell cycle regulating proteins was analyzed by SDS-PAGE followed by Western blotting using the indicated specific antibodies. Blotting of Actin was used as a loading control. B. Schematic representation of the human Cdc25B promoter regions, indicating sites for Sp1, Sp3, and NF-YB motifs, amplified from the precipitated DNA by specific primer sets 1 and 2. C. ChIP analysis for IER5 in the Cdc25B promoter. D & E. IER5-dependent changes in Sp1, Sp3 and NF-YB binding (E), and the bindings of coactivators, DNMT1 and p300, to interact with Sp1 and NF-YB, respectively (F) in Cdc25B promoter. U937 cells were either untransfected or were transfected with the control DNA or IER5 DNA. After 3 days culture, cells were cross-linked with 1% formaldehyde and ChIP was performed with either control antibody (IgG) or the anti-IER5 antibody. The precipitated DNA was then assayed by PCR using pairs (primer set 1 and 2) of oligonucleotides encompassing specific regions of the Cdc25B promoter. The values of ChIP efficiencies are given as % of input. doi:10.1371/journal.pone.0028011.g004 G2/M phase as well as in S phase compared to control untreated cells, 3 days after transfection. These data are consistent with the established notion that IER5 is a key regulator of cell cycle checkpoints [11].
Because induction or over-expression of IER5 expression induced G2/M cell cycle arrest and inhibited AML cell proliferation, we considered that identification of genes regulated by IER5 would facilitate an understanding of the roles IER5 plays in AML cell proliferation. We previously showed that knock-down of FOXM1, or TMPP treatment, markedly decreased the protein expression of the mitotic regulators Skp2, Cdc25B, Cyclin B1, Survivin, KIS, Aurora-B kinase in AML cells [1,2]. Diminished expression of these mitotic regulators results in a block in mitotic progression or proliferation, and leads to an accumulation of p27 Kip1 and p21 Cip1 proteins [1,2]. Moreover, the expression of Cdc25B is known to be increased in AML cells [13]. In this study, we found that IER5 over-expression markedly decreased Cdc25B protein expression, but not the expression of CHK1, WEE1, Cyclin B1, Survivin, and Aurora kinase B proteins. Thus, of the genes in the G2/M specific gene cluster, among which are essential regulators of mitosis, IER5 predominantly controls the expression of Cdc25B. Therefore, IER5 induction would be expected to strongly induce G2/M cell cycle arrest in AML cells. In addition, we observed the number of cells in both S-and G2/ M-phase. In cell cycle, the G2/M checkpoint is a barrier to aberrant cellular proliferation, and arrests damaged cells to allow time for repair [17]. The expression of Cdc25B is initiated during S-phase, peaking in G2 and declining in mitosis [18]. Cdc25B has been reported to play a role in the control of entry into mitosis after a DNA damage-activated checkpoint in cells [14,19,20]. In AML cells, the Cdc25B participates in the G2/M checkpoint recovery and its expression is upregulated, and the reduction of Cdc25B expression induced the cell number of both S-and G2/ M-phase in AML cells [14]. Our data are consistent with the data of the report [14]. Therefore, when TMPP induced the IER5 expression, the Cdc25B expression is reduced and G2/M checkpoint is induced, and then the cells with S-and G2/Mphase DNA contents is increased. Moreover, the population of G2/M distribution is more significantly increased than that of Sphase.
Cdc25 phosphatases, which are dual specificity phosphatases and important regulators of G2/M transition regulators, regulate entry into mitosis by regulating the activation of CDK1/cyclin B [21]. The Cdc25 phosphatases activate the mitotic CDK complexes by dephosphorylating the inhibitory Thr14 and Tyr15 residues on CDK1 and CDK2 [22]. There are three isoforms of the Cdc25 phosphatase; Cdc25A, Cdc25B and Cdc25C, and all three isoforms have roles in G2/M progression. It was previously believed that Cdc25A acts at the G1/S transition, whereas Cdc25B and Cdc25C function mainly at the G2/ M transition [21,23]. However, recent studies suggest that all three CDC25 phosphatases function as regulators of both G1/S and G2/M transitions [23]. It has also been reported that only depletion of Cdc25A and Cdc25B delays entry into mitosis [24]. In contrast, all three Cdc25 isoforms are targets for checkpoint kinase inactivation [25]. Cdc25A is destabilized by CHK1 phosphorylation in response to DNA damage [26]. Cdc25B is specifically required for exit from the G2 phase checkpoint arrest [27], and its stability has also been linked to responses to damage [28]. Thus, it is believed that aberrant regulation of Cdc25 phosphatases causes tumorigenesis via dysregulation of the cell cycle [22,[29][30][31]. It has been suggested that Cdc25 phosphatases are involved in deficient checkpoints during mitosis in malignant transformation [22,32]. The activity of the Cdc25 phosphatases is regulated by their phosphorylation status, their expression level and their subcellular localization [22,32]. Abnormal expression of Cdc25B has been reported in a number of carcinomas, such as non-small cell lung cancer [29], colorectal carcinomas [32], ovarian [33], esophageal [34,35], prostate [36], gastric [37] and pancreatic [38] cancers. Thus, overexpression of Cdc25B is believed to contribute to tumorigenesis by enhancing tumor malignancy [21]. However, it remains unknown how Cdc25B expression is translationally regulated. Currently, it has been reported that Cdc25B is negatively regulated by p53 through Sp1 and NF-Y transcription factors [14]. The analysis of Cdc25B promoter shows that the down-regulation of Cdc25B by p53 is needed for the Sp1/Sp3 and NF-Y binding sites in Cdc25B promoter, p53 binds to the Cdc25B promoter and mediates transcriptional attenuation via the Sp1 and NF-Y transcription factors. We have examined the relationship between IER5 and Cdc25B promoter, and binding of Sp1 and NF-YB, which is one of NF-Y subunits, by using ChIP assay in AML cell lines and AMLderived ALDH hi /CD34 + cells. In leukemia cells, IER5 directly binds to Cdc25B promoter. Moreover, when IER5 over-expressed, the significantly reduced binding of NF-YB on the Cdc25B promoter the release of anti-histone acetyltransferase p300, which is known as a coactivator of NF-Y [15], was observed at upstream of 1st exon of Cdc25B. On the other hand, we did not observe that Figure 5. IER5 inhibits the colony formation of ALDH hi /CD34 + cells isolated from AML patients. A. Selection of ALDH hi /CD34 + hematopoietic progenitor cells from the bone marrow of two healthy volunteers (#1 and #2) by FACS sorting. Region P denotes populations of ALDH hi cells. Region R and S denote populations of CD34 + and CD34cells in the ALDH hi population (Region P), respectively. Negative control, light grey region; CD34-PE staining, dark grey region. B. Selection of ALDH hi /CD34 + hematopoietic progenitor cells from the bone marrow of two AML patients (M1 and M2) by FACS sorting. Region P denotes populations of ALDH hi cells. Region R and S denote populations CD34 + and CD34cells in the ALDH hi population (Region P), respectively. Negative control, light grey region; CD34-PE staining, dark grey region. C. ALDH hi /CD34 + cells were purified from a healthy volunteer (#1) and an AML patient (M1), and were cultured in semisolid methylcellulose media. The ALDH hi /CD34 + cells from each source were left untransfected or were transfected with IER5 cDNA, or treated with TMPP (5 mM). After 14 days culture, the colony forming ability of the cells was analyzed (left upper panel) and the cells were viewed using phase-contrast microscopy. Original magnification 64 (left bottom panels). Their mRNA expression of IER5 and Cdc25B was assessed using RT-PCR and quantitative RT-PCR (right panels). Colonies formed by these ALDH hi /CD34 + cells (3610 2 to 5610 2 cells/plate) were counted following plating in semisolid methylcellulose media. Colony formation was evaluated by determination of colony counts as a percentage of the corresponding control. The results are the means 6 SD of three independent experiments. *P,0.01 compared with untreated control cells. RT-PCR results representative of three independent experiments are shown. GAPDH mRNA expression is shown as an internal control. The ALDH hi /CD34 + cells whose IER5 mRNA expression was analyzed by quantitative RT-PCR were derived from an AML patient (M1). The levels of the quantified RT-PCR products were normalized to GAPDH expression in the same sample and were then expressed relative to the mRNA level of a normal control which was assigned a value of 1. D and E. Cell cycle analysis (D) and changes of mitochondrial membrane potential (DYm) (E) in IER5 overexpressed or TMPP treated AML-derived ALDH hi /CD34 + cells. ALDH hi /CD34 + cells were purified from an AML patient (M1), and were then transfected with IER5 cDNA or were treated with TMPP (5 mM). The IER5-transfected or TMPPtreated cells were harvested after 3 days. The cell cycle distribution and the DYm of the ALDH hi /CD34 + cells was analyzed using flow cytometric analysis. The FACS results are representative of three independent experiments. NC; Negative control. doi:10.1371/journal.pone.0028011.g005 IER5 interacts with Sp1 in AML cells. These results demonstrate that the binding of IER5 on the Cdc25B promoter induced the recruitment of coactivator p300, causing the inhibition of AML cell proliferation and colony formation of AML-derived ALDH hi / CD34 + cells through the down-regulation of Cdc25B.
We performed a colony formation assay using the ALDH hi / CD34 + progenitor cells derived from two normal healthy volunteers and two AML patients (M1 and M2). The number of colonies was moderately reduced when normal ALDH hi /CD34 + progenitor cells were transfected with IER5 cDNA. In contrast, colony numbers were significantly reduced when ALDH hi /CD34 + AML progenitor cells were transfected with IER5 cDNA or were treated with TMPP. Moreover, the relative expression of Cdc25B was decreased in colony forming cells when ALDH hi /CD34 + AML progenitor cells were treated with TMPP or were transfected with IER5 cDNA. Moreover, when ALDH hi /CD34 + AML progenitor cells were transfected with IER5 shRNA, the inhibitory effects of TMPP on colony formation by TMPP were decreased in colony forming cells. These results indicate that increased expression of IER5 induced by TMPP reduced the number of colony forming cells derived from AML progenitor cells through a reduction in Cdc25B expression. Consistent with the idea that IER5 and TMPP effects were mediated by a decrease in Cdc25B, we found that over-expression of Cdc25B rescued both the inhibitory effects of TMPP treatment and those of IER5 overexpression in ALDH hi /CD34 + AML progenitor cells, resulting in recovery of the colony forming activity of AML-derived ALDH hi / CD34 + cells.
In conclusion, this study shows for the first time that IER5 overexpression strongly inhibited AML-derived ALDH hi /CD34 + cell proliferation and colony formation through a reduction in Cdc25B expression, resulting in induction of G2/M cell cycle arrest. Moreover, our data show that the reduction of NF-YB by IER5 inhibited the AML cell proliferation through the release of p300. Therefore, IER5 induction could be an attractive potential target for AML therapy.

Materials and Methods
Chemical synthesis and reagents TMPP was prepared as previously described by reacting 4bromo-3-methyl-1-phenyl-2-phospholene 1-oxide with bromine [1]. The reaction mixtures were extracted with chloroform, washed with saturated NaCl solution, and dried with anhydrous sodium sulfate. These reaction agents were then dissolved in dimethyl sulfoxide (DMSO) (Sigma-Aldrich, St Louis, MO) and diluted in culture medium immediately before use. The final concentration of DMSO in all experiments was less than 0.01%, and all treatment conditions were compared with vehicle controls. Cytarabine (cytosine arabinoside, Ara-c) was purchased from Sigma-Aldrich.

Cells and cell cultures
The human leukemia cell lines, KG-1, Kasumi-1 and U937, were purchased from the American Type Culture Collection (ATCC, Manassas, VA). YRK2 cells were harvested in our Figure 6. IER5 expression inhibited the colony formation of AML ALDH hi /CD34 + cells through the regulation of Cdc25B expression. A. ALDH hi /CD34 + cells were purified from an AML patient (M1) and were cultured in semisolid methylcellulose media. The ALDH hi /CD34 + cells were either untransfected or were transfected with IER5 cDNA or with shRNA-#1, or -#2. After 14 days culture with or without TMPP (5 mM), the cells were then analyzed for their colony forming ability. Colony formation of these ALDH hi /CD34 + cells (3610 2 to 5610 2 cells/plate) was assessed after plating in semisolid methylcellulose media. B. Analysis of the mRNA expression of IER5 and Cdc25B in each colony using QRT-PCR and RT-PCR. C. The cells were viewed using phase-contrast microscopy after 14 days culture with or without TMPP (5 mM). Original magnification 64. D. ALDH hi /CD34 + cells were purified from an AML patient (M1) and were cultured in semisolid methylcellulose media. The ALDH hi /CD34 + cells were either untransfected or were transfected with IER5 cDNA or Cdc25B cDNA. After 14 days culture with or without TMPP (5 mM), the cells were then analyzed for their colony forming ability. Colony formation of these ALDH hi /CD34 + cells (3610 2 to 5610 2 cells/plate) was assessed after plating in semisolid methylcellulose media. Colony formation was evaluated as a percentage of the corresponding control. E. Analysis of the mRNA expression of IER5 and Cdc25B in each colony using QRT-PCR and RT-PCR. The levels of the QRT-PCR products were normalized to GAPDH expression in the same sample and were then expressed relative to the mRNA level of a normal control which was assigned a value of 1. RT-PCR results representative of three independent experiments are shown. GAPDH mRNA expression is shown as an internal control. The results are the means 6 SD of three independent experiments. *P,0.01 compared with untreated control cells. F. The cells transfected with Cdc25B or IER5 cDNA were viewed using phase-contrast microscopy after 14 days culture with or without TMPP (5 mM). Original magnification 64. doi:10.1371/journal.pone.0028011.g006 laboratory from bone marrow samples of AML (French-American-British (FAB) classification; M5a) patients after obtaining informed consent. The KG-1, U937, and YRK2 cells were cultured in RPMI 1640 media containing 10% heat-inactivated fetal bovine serum (FBS), 2 mM L-glutamine, 100 mg/ml streptomycin, and 200 U/ml penicillin (GIBCO-BRL, Gaithersburg, MD). Kasumi-1 cells were grown in RPMI1640 containing 20% FBS, 2 mM L-glutamine, 100 mg/ml streptomycin, and 200 U/ml penicillin. Primary leukemia cell specimens were obtained from the bone marrow of two AML patients (FAB classification; M1 and M2) before the start of any treatment, and normal hematopoietic cells were extracted from healthy donors (n = 2) after obtaining informed consent, which was written. We obtained ethics approval for this study from the Institutional Review Board of Hamamatsu University School of Medicine. Mononuclear cells (MNCs) were purified by Ficoll-Hypaque density-gradient centrifugation. The ALDH hi cells were gated, and the CD34 + cells in the gated ALDH hi population were sorted on a fluorescence-activated cell sorter (Becton Dickinson). Sorted ALDH hi /CD34 + cell populations were collected in methylcellulose media (Methocut H4435; StemCell Technologies, Inc., Vancouver, BC, Canada).

Plasmids and RNA interference
Full-length cDNAs encoding human IER5 and Cdc25B were obtained by RT-PCR using human bone marrow cDNA (BD Biosciences Clontech, Palo Alto, CA) as a template and were cloned into the eukaryotic expression vector pcDNA3.1/V5-His (Invitrogen, Carlsbad, CA). Sequences of recombinant IER5 and Cdc25B cDNAs were verified using automated sequencing.
The vectors for RNA interference (RNAi) specific for human IER5 were constructed based on the piGENE PUR hU6 vector (iGENE Therapeutics, Tsukuba, Japan), according to the manufacturer's instructions. We used the following targeting sequences: IER5 shRNA #1; 59-CCTCATCAGCATCTTCGGT-39, and IER5 shRNA #2; 59-CTGCATAAGAACCTCCTG-39. The scrambled shRNA sequence, 59-GGACGAACCTGCTGAGA-TAT-39, was used as a control. Vectors were transfected into cells by using the Lipofectamine 2000 kit (Life Technologies, Gaithersburg, MD), according to the manufacturer's instructions. The transfection procedure was repeated 12 h after the first transfection, and cells were harvested at 48 and 72 h after the initial transfection. Knockdown efficiency was consistently 60% to 70%, as determined by RT-PCR measurement of IER5 mRNA.
Cell proliferation and viability assay KG-1, Kasumi-1, U937 and YRK2 cells were transfected with IER5 shRNA-#1 or -#2, with scrambled shRNA, or were left untreated. After 3 days incubation, the cells were plated in six-well plates at a density of 1610 4 cells per well. Cell proliferation was measured by counting cells using a hemocytometer. AML cells were seeded in 96-well, flat-bottom microplates at a density of 2610 4 cells per well. The cells were untransfected or were transfected with scrambled shRNA or with IER5 shRNA-#1 or -#2. Cells grown in complete medium without transfection were used as controls. Cell proliferation was also assessed by counting of viable cells on the indicated days using trypan blue (Sigma-Aldrich) exclusion. The number of nonviable cells was determined by counting of live cells that did not uptake trypan blue in a hemocytometer and are reported as the percentage of untransfected control cells. Each data point was performed in triplicate, and the results are reported as mean counts 6 SD.

Detection of changes in mitochondrial membrane potential (DYm)
To detect DYm, cells (1610 4 cells/well) that were transfected with IER5 cDNA or treated with TMPP were incubated in 24-well plates. After 3 days, cells were labeled with DiOC6 (40 nM in culture medium) at 37uC for 20 min. After washing in PBS, cellular uptake of DiOC6 was analyzed using flow cytometry.

Cell cycle analysis
Cellular DNA content was analyzed using propidium iodide (PI) (Sigma-Aldrich) staining. The cells were stained with 50 mg/ml PI on day 3 post-transfection. The relative DNA content per cell was measured using flow cytometry and an Epics Elite flow cytometer (Coulter Immunotech, Marseille, France). The percentage of cells in G1, S, and G2/M phases was calculated using the ModFit program (Becton Dickinson, San Jose, CA).

Western blot analysis
Cells transfected with scrambled shRNA or with IER5 shRNA-#1 or -#2 were harvested after 3 days. Western blot analysis was performed using the following antibodies: goat polyclonal anti-IER5 (Abcam, Cambridge, UK); Rabbit polyclonal anti-Cdc25B, anti-CHK1, anti-WEE1 and anti-Aurora-B; mouse monoclonal anti-Cyclin B1 and anti-Survivin, all from Santa Cruz. To ensure equal protein loading, Western blotting of actin, used as an internal control, was carried out using a mouse monoclonal anti-Actin antibody (C-4; ICN, Aurora, OH).

Colony forming assays
Colony forming assays were performed by plating purified populations of cells at concentrations ranging from 1610 3 to 2610 3 cells into methylcellulose media (Methocut H4435; Stem Cell Technologies). Colonies were enumerated under light microscopy (Zeiss, Munchen, Germany) following incubation for 14 days at 37 uC and 5% CO 2 .

Isolation of progenitor cells and QRT-PCR of progenitor cells
Following colony formation, each colony was harvested using a glass syringe, and all cells of the same colony were pooled and washed. An RNeasy system was used to extract total RNA from approximately 5610 4 cells from each colony.

Statistical analysis
Data are representative of at least three experiments with essentially similar results. These results are expressed as the means 6 standard deviations (SD) or standard error of the means as indicated. Statistical analyses of the data were performed using Student's t-test. P values less than 0.05 were considered statistically significant.