Vav1 is necessary for PU.1 mediated upmodulation of miR‐29b in acute myeloid leukaemia‐derived cells

Abstract It has been recently demonstrated that high pre‐treatment levels of miR‐29b positively correlated with the response of patients with acute myeloid leukaemia (AML) to hypomethylating agents. Upmodulation of miR‐29b by restoring its transcriptional machinery appears indeed a tool to improve therapeutic response in AML. In cells from acute promyelocytic leukaemia (APL), miR‐29b is regulated by PU.1, in turn upmodulated by agonists currently used to treat APL. We explored here the ability of PU.1 to also regulate miR‐29b in non‐APL cells, in order to identify agonists that, upmodulating PU.1 may be beneficial in hypomethylating agents‐based therapies. We found that PU.1 may regulate miR‐29b in the non‐APL Kasumi‐1 cells, showing the t(8;21) chromosomal rearrangement, which is prevalent in AML and correlated with a relatively low survival. We demonstrated that the PU.1‐mediated contribution of the 2 miR‐29b precursors is cell‐related and almost completely dependent on adequate levels of Vav1. Nuclear PU.1/Vav1 association accompanies the transcription of miR‐29b but, at variance with the APL‐derived NB4 cells, in which the protein is required for the association of PU.1 with both miRNA promoters, Vav1 is part of molecular complexes to the PU.1 consensus site in Kasumi‐1. Our results add new information on the transcriptional machinery that regulates miR‐29b expression in AML‐derived cells and may help in identifying drugs useful in upmodulation of this miRNA in pre‐treatment of patients with non‐APL leukaemia who can take advantage from hypomethylating agent‐based therapies.

down-regulated in AML and whose restoration into AML cell lines or primary samples led to a dramatic reduction of tumorigenicity. [5][6][7] High pre-treatments levels of miR-29b have been recently associated with longer survival in AML patients treated with conventional chemotherapy and to improved clinical response to DNA methyltransferases (DNMT) inhibitors, 8,9 suggesting that strategies aimed to increase this miRNA may be useful in therapeutic DNA hypomethylation of leukaemic blasts. As downmodulation of miR-29b correlates with different pathologies, a number of delivery systems for exogenous miR-29b have been generated. [10][11][12][13] Concerning AML, a transferrin-conjugated nanoparticle method was developed to increase miR-29b in AML blasts. 14 However, despite its efficacy in in vivo AML models, this approach is not clinically available at present and activation of transcriptional expression may constitute an efficient therapeutic strategy for restoring the miR-29b level in myeloid leukaemia.
The precursors of the miR-29 family are transcribed in 2 clusters: a miR-29a/b1 cluster located on chromosome 7 (7q32) and a miR-29b2/c cluster located on chromosome 1 (1q32). The 2 distinct precursor sequences encoding for miR-29b lead to identical mature products. 10 Different binding sites for several transcriptional factors have been identified in the promoter of miR-29a/b1 and miR-29b2/ c clusters in different tissues. 5 In myeloid leukaemia cells, CEBPa was reported to only regulate the miR-29a/b1 cluster, providing the rationale for miR-29b suppression in AML patients with loss of chromosome 7q or deficiency of this transcription factor. 15 The miR-29b2/c locus was shown to be activated at transcriptional level by the master myeloid regulator PU.1 in cells from acute promyelocytic leukaemia (APL) treated with all-trans-retinoic acid (ATRA). 16 PU.1 is generally deregulated in AML by mechanisms including interference with binding sites (including at its own promoter site) by PML-RARA 17 or disruption of PU.1 transactivation activity by RUNX1-ETO (AML1-ETO). 18 Even if a number of differentiating agonists may restore the levels of PU.1 in AML-derived cells, 19,20 at present, only APL patients are treated and take advantage by the use of differentiating therapies. 1,2 A crucial element of PU.1 transcriptional activity results from its ability to interact with a number of different protein partners, whose identification may provide the starting point for the development of targeted therapies for the treatment of haematological malignancies. 21 In APL-derived cells treated with ATRA, the interaction of PU.1 with its recognition sites on the CD11b 22 and miR-142-3p 23 promoters is entirely dependent from adequate amount of Vav1, a multidomain protein involved at different levels in agonist-induced differentiation of APL-derived promyelocytes. 24 In addition, Vav1 may regulate protein expression of ATRA-treated APL-derived cells by determining the nuclear amount of proteins implicated in mRNA production and stability. 25 This work was first aimed to assess the role of PU.1 in regulating the expression of miR-29b in non-APL cells, in order to identify agonists that, through upmodulation of this transcription factor, could be beneficial in DNA hypomethylation-based therapies. We used Kasumi-1 cells, showing the t(8;21) chromosomal translocation, that represents the most common cytogenetic subtype of AML. In this cell line, the ectopic expression of PU.1 overcomes its functional block induced by AML1-ETO, in turn involved in a regulatory circuit with miR-29b1 that controls the leukaemic phenotype. 18,26 As we demonstrated that, in APL-derived cell treated with ATRA, Vav1 is crucial for the interaction of PU.1 with its DNA consensus regions on miR-142 promoter, the cooperation between the 2 proteins in modulating miR-29b expression was investigated in both APL-and non-APL-derived cells.

| MATERIALS AND METHODS
All reagents were from Sigma Chemicals Co. (St Louis, MO, USA) unless otherwise indicated.

| Cell culture and treatments
The human myeloid leukaemia Kasumi-1 (t(8;21)) and the APL- France) at 37°C in a humidified atmosphere containing 5% CO 2 in air. The cell density was maintained between 5 9 10 5 /mL and 1.5 9 10 6 /mL. Cells were monthly tested for mycoplasma and other contaminations and quarterly subjected to cell identification using single nucleotide polymorphism (SNP) typing.
To establish the percentage of adherent cells, after removal of cells in suspension, the cells adhering to the flask were detached with a trypsin/EDTA solution (Gibco Laboratories). Both suspended and adherent cells were counted using a hemocytometer, and the level of adhesion was expressed as a percentage of adherent cells over the total number of cells.

| Immunoprecipitation and Western blot analysis
Total lysates from both Kasumi-1 and NB4 cells were obtained by adding Laemmli's SDS sample buffer to cells, after washing with cold PBS containing 1 mmol/L Na 3 VO 4, .
Purification of nuclei from both NB4 and Kasumi-1 cells was performed essentially as previously reported, 22 with the only modification consisting in the use of a 20-gauge needle for Kasumi-1.
For immunoprecipitation experiments, nuclei from ATRA-treated NB4 and PMA-treated Kasumi-1 were lysed, added of protease and phosphatase inhibitors, were incubated with antibodies directed against Vav1 or PU.1 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and immunoprecipitated with protein A-Sepharose (Pharmacia, Uppsala, Sweden), essentially as previously reported. 22 For Western blot analysis, total cell lysates and immunoprecipitates from nuclei were separated on 8.5% polyacrylamide denaturing gels and blotted to nitrocellulose membranes (GE Healthcare Life Science, Little Chalfont, UK). The membranes were then reacted with antibodies directed against PU.1 and Vav1 (Santa Cruz Biotechnology) and against b-tubulin (Sigma), incubated with peroxidase-conjugated secondary antibodies and revealed using the ECL system (PerkinElmer, Boston, MA, USA), as previously reported. 22 The chemiluminescence-derived bands were acquired with ImageQuant TM LAS 4000 biomolecular imager (GE Healthcare), and the densitometrical analysis was performed by means of Image Quant TL software (GE Healthcare).

| RNA interference assays
Exponentially growing Kasumi-1 and NB4 cells were transfected with a mixture of small interfering RNAs (siRNAs; Santa Cruz Biotechnology) targeting the mRNAs for PU.1 or Vav1, using a previously described electroporation procedure. 22,27 As a control for transfection efficiency, which was always higher than 60%, a nonsilencing fluorescein-labelled duplex RNA (Qiagen S.p.A, Milan, Italy) was used. 5 hours after transfection, cells were treated with ATRA or PMA, incubated at 37°C in a 5% CO 2 atmosphere and then subjected to immunochemical or immunocytochemical analysis, to miR-29b evaluation and to quantitative chromatin immunoprecipitation experiments.

| Quantitative Real-time PCR assay (qRT-PCR)
High-quality small RNAs from Kasumi-1 and NB4 cells were extracted using a miRNeasy Micro Kit (Qiagen) as previously reported. 23 Briefly, 10 ng RNA was subjected to single-stranded cDNA synthesis, and the obtained cDNAs were employed as templates for quantitative Real-time PCR-based miR-29b expression measurements using TaqMan MicroRNA Assays (ID 000413; Life Technologies). Thermal cycling and fluorescence detection were performed according to the manufacturer's instructions, using a Bio-Rad CFX96 TM sequence detection system (Bio-Rad Laboratories, Hercules, CA, USA), and the data were analysed using a dedicated software (Bio-Rad Laboratories). miRNA expression levels were normalized to U6 snRNA (Life Technologies), and fold change was determined using the 2 ÀDDCt method. Cycle threshold >35 was excluded. Control PCR samples were run without cDNA. All reactions were performed in triplicate, and the experiments were repeated 3 times.

| Quantitative chromatin immunoprecipitation (Q-ChIP) assay
Quantitative chromatin immunoprecipitation experiments were performed on untreated and treated Kasumi-1 and NB4 cells using a ChIP assay kit (Upstate Biotechnology, Lake Placid, NY, USA) as previously reported. 23,27 Samples were subjected to immunoprecipitation at 4°C overnight with antibodies directed against PU.1 or Vav1 or with a non-specific IgG, used as a negative control (Santa Cruz Biotechnology). Beads were then washed, protein/DNA complexes eluted and cross-links reversed by heating samples at 65°C overnight. After protein digestion, DNA was recovered using a PCR purification kit (Promega, Madison, WI, USA) in 50 lL elution buffer.
Quantitative PCR of (i) a 170-bp DNA fragment, encompassing the putative PU.1 binding site located at À330/À324 bp from the transcriptional start in the human primiR-29a/b1 promoter on chromosome 7q32.3 and of (ii) a 181-bp DNA fragment, encompassing the putative PU.1 binding site located in the proximal miR-29b2/c promoter on chromosome 1q32.2, was performed in triplicate using an iTaq Universal SYBR green SuperMix on a Bio-Rad CFX96 TM Real-time detection system (Bio-Rad Laboratories). The primers used were as follows: (i) Fw: 5 0 -GCAGAGGATTAGACAGAGGGTG-3 0 , Rev: ChIP-qPCR data are presented as relative to input signals and in comparison with the background signals (IgG). PCR products were separated on tris-acetate 1% agarose gels, stained with ethidium bromide and visualized by UV light apparatus.

| Statistical analysis
Statistical analysis was performed using the 2-tailed Student's t test for unpaired data with the GraphPad Prism 6.0 statistical package (GraphPad Software, San Diego, CA, USA). P values < .05 were considered statistically significant.  18 On the basis of the described PU.1-mediated expression of miR-29b induced by ATRA in NB4 cells, 16 this agonist was first administered to Kasumi-1. As expected, 29 ATRA induced a slight cell adhesion ( Figure S1A), indicative of a partial differentiation along the monocyte-macrophage lineage, and a substantial increase in PU.1 ( Figure 1A,B). Nevertheless, unlike what was observed in NB4 cells, 16 ATRA failed to significantly upmodulates the miR-29b levels in Kasumi-1 ( Figure 1C, Figure S1B).
We then treated Kasumi-1 with PMA, known to activate PU.1 in myeloid cells. 30 We found that PU.1 expression increased in treated conditions ( Figure 1A For this purpose, ChIP assay was performed using specific primers that amplify the potential PU.1 binding sites in the 5 0 regulatory regions ( Figure 2A). As shown in Figure 2B, PU.1 was selectively recruited to the miR-29b locus on chromosome 7 also in control conditions and PMA, but not ATRA, led to a significant increase in DNA associated with the transcription factor. A region of DNA flanking the miR-29b promoter on Chr 7 and not predicted to bind PU.1 was also used in the ChIP assays, confirming the specific in vivo recruitment of this transcription factor in both control and PMA-treated Kasumi-1 ( Figure S2) 3.2 | In NB4 cells, Vav1 is essential for binding of PU.1 to its consensus sequences located on the miR-29b promoters As we previously found that, in APL-derived cells, Vav1 regulates the presence of PU.1 on its consensus region on the miR-142-3p promoter, 23 our subsequent aim was to assess whether, also in regulation of miR-29b, the PU.1 action is supported by Vav1. We first addressed this issue in NB4 cells treated with ATRA, in which, as expected, 22 a significant increase in both PU.1 and Vav1 levels was induced by the agonist (Figure 3A,B). As expected, 16 16 we investigated here the ability of this transcription factor to modulate miR-29b in non-APL cells, in order to identify agonists that, up-regulating this transcription factor may be useful in hypomethylation-based therapies. As experimental model, we choose the myeloid-derived Kasumi-1 cells, displaying the t(8;21) chromosomal rearrangement, the most common cytogenetic subtype of AML, whose survival rate is as low as 30% on a 5-year basis. 3 In Kasumi-1, the notion that overexpression of PU.1 overcomes its functional block induced by the fusion protein AML1-ETO, in turn involved in a phenotype regulatory circuit with miR-29b, 26