B-Cell Regulator of Immunoglobulin Heavy-Chain Transcription (Bright)/Arid3a Is a Direct Target of the Oncomir Microrna-125B in Progenitor B-Cells

ORIGINAL ARTICLE B-cell regulator of immunoglobulin heavy-chain transcription (Bright)/ARID3a is a direct target of the oncomir microRNA-125b in progenitor B-cells

MP Puissegur1,6, R Eichner1,6, C Quelen1, E Coyaud1, B Mari2, K Lebrigand2, C Broccardo1, F Nguyen-Khac3,4, M Bousquet1 and P Brousset1,5

B-cell acute lymphoblastic leukemia (B-ALL) is often associated with chromosomal translocations leading to the deregulation of proto-oncogenes. MicroRNAs can also be affected by chromosomal alterations and thus contribute to carcinogenesis. The microRNA, miR-125b-1, is overexpressed in B-ALL cases with the t(11;14)(q24;q32) translocation; therefore, we sought to determine the role of this microRNA in B-cell fate. We used murine pre-BI cells alongside murine and human leukemic B-cell lines to show that miR-125b expression enhances proliferation by targeting B-cell regulator of immunoglobulin heavy-chain transcription (Bright)/ ARID3a, an activator of immunoglobulin heavy-chain transcription. Accordingly, this target gene was downregulated in B-ALL patients with the t(11;14)(q24;q32) translocation. Repression of Bright/ARID3a blocked differentiation and conferred a survival advantage to Ba/F3 cells under interleukin-3 starvation. In addition, overexpression of miR-125b protected pre-BI and leukemic B-cell lines from apoptosis by blockade of caspase activation by a mechanism that was independent of p53 and BAK1. In summary, miR-125b can act as an oncogene in B-ALL by targeting ARID3a and mediating its repression, thus leading to a blockage in differentiation, increased proliferation and inhibition of apoptosis.

Leukemia (2012) 26, 2224–2232; doi:10.1038/leu.2012.95 Keywords: microRNA; B-cell acute lymphoblastic leukemia; B-cell differentiation; proliferation

INTRODUCTION cases, miR-125b-1 seems to contribute to oncogenesis by blocking 5 B-cell acute lymphoblastic leukemia (B-ALL) is characterized by a myeloid differentiation. miR-125b is also involved in B-cell population of immature lymphoblastic cells with unlimited malignancies such as B-ALLs containing the ETV6/RUNX1 proliferative potential and arises from progenitor B-cells at the translocation, although in these cells miR-125b-2 overexpression is independent of the ETV6/RUNX1 fusion protein and is thought pro/pre-B-cell stage. Two distinct events occur during malignant 7 transformation: (1) a blockage of differentiation at the pre-B-cell to confer a survival advantage. Three other studies have reported stage and (2) an acquisition of unlimited proliferative potential.1 B-ALL cases with an isolated chromosomal aberration whereby miR-125b-1 is found inserted into the immunoglobulin heavy- The primary events leading to malignant transformation are often 8–10 structural mutations such as deletions, ampliﬁcations or chain locus, resulting in its upregulation. translocations that result in either the generation of fusion Given these cases where miR-125b is involved in B-ALL and the fact that miR-125b blocks differentiation in myeloid progenitor genes or the deregulation of proto-oncogenes. As well as 5 classical protein-coding genes, microRNAs are often upregulated cells, we sought to determine the effects of miR-125b or downregulated by such chromosomal translocations as they are overexpression in B-cell progenitors and in B-ALL. frequently found at fragile sites within the genome.2,3 miR-125b, the homolog of the Caenorhabditis elegans lin-4 gene, is one of the microRNAs most frequently associated with MATERIALS AND METHODS cancer.4–8 In humans, there are two miR-125b gene loci: miR- Patient samples 125b-1 on chromosome 11q23 and miR-125b-2 on chromosome Samples of complimentary DNA from B-ALL patients with and without the 21q21; these are differentially expressed depending on the tissue chromosomal translocation t(11;14)(q24;q32) were provided by Dr Florence N’Guyen-Khac. Clinical details are already published, as type. Although their pri-microRNAs have slightly different 9 sequences, mature miR-125b-1 and miR-125b-2 are identical described in Chapiro et al. and both are reported to be overexpressed in certain solid 4 5,6 tumors and hematological malignancies. In hematological Cell culture malignancies, miR-125b-1 has been shown to be overexpressed The Phoenix retrovirus producer line (Orbigen, Inc., San Diego, CA, USA) in some cases of acute myeloid leukemia and myelodysplastic was cultured in Dulbecco’s Modiﬁed Eagle’s Medium—high glucose (Life syndrome secondary to a t(2;11)(p21;q23) translocation. In these Technologies, Foster City, CA, USA), 10% fetal calf serum (FCS) (Invitrogen,

1Institut National de la Sante´ et de la Recherche Me´dicale, U1037, Centre de Recherches en Cance´rologie de Toulouse, 31300 Toulouse, France; Universite´ Paul Sabatier, Toulouse, France; 2Institut de Pharmacologie Molećulaire et Cellulaire, CNRS UMR6097, Sophia Antipolis, France; 3Service d’He´matologie Biologique, Hoˆpital Pitie´-Salpeˆtrie`re, Paris 6, France; 4INSERM U872, UPMC Paris 6, France and 5De´partement de Pathologie, Centre Hospitalier Universitaire Purpan, Toulouse, France. Correspondence: Dr P Brousset, De´partement de Pathologie, Centre Hospitalier, Universitaire Purpan, Place Baylac, Toulouse Cedex 31059, France. E-mail: [email protected] 6These authors contributed equally to this work. Received 29 July 2010; revised 12 March 2012; accepted 22 March 2012; accepted article preview online 3 April 2012; advance online publication, 4 May 2012 miR-125b targets ARID3a in B-ALL MP Puissegur et al 2225 Carlsbad, CA, USA). Murine OP9 bone marrow stromal cells were cultured in minimum essential medium-a (Life Technologies), 20% FCS. Murine wild-type pre-BI cells were isolated from C57BL/6 mouse bone marrow as being B220 þ /cKIT þ /Igk À and were cultured on a monolayer of irradiated OP9 stromal cells in Iscove’s Modified Dulbecco’s Medium (Life Technol- ogies), supplemented with 20% FCS, 100 units/ml murine interleukin (IL)-7 (Peprotech, RockyHill, CT, USA), 0.05 mM b-Mercaptoethanol (Sigma- Aldrich, St Louis, MO, USA) and 2 mML-Glutamine (Invitrogen). Every 3 days, pre-BI cells were harvested and propagated on fresh OP9 stromal cells. Murine B-ALL cell lines 38B9 and 18–81, and human Nalm6 pre-B- cells were cultured in RPMI (Life Technologies), 10% FCS, 2 mML-Glutamine and 1 mM Sodium-Pyruvate (Invitrogen). This was supplemented with either 0.05 mM b-Mercaptoethanol or 100 units/ml murine IL-3 (Peprotech) to culture the murine B-ALL cell lines 70Z/3 and BaF/3, respectively. All cells were cultured in the presence of penicillin (100 U/ml) and streptomycin (100 mg/ml) at 37 1C in a 5% CO2 atmosphere. Figure 1. MicroRNA expression was determined by qRT-PCR using Taqman MicroRNA assays (Life Technologies, Foster City, CA, USA). Antibodies and cell sorting The y axis represents relative expression levels of mature miR-125b normalized using an internal control (sno202). For fluorescence-activated cell sorting (FACS) and flow cytometry analyses, we used the following antibodies: phycoerythrin/Cy5.5-conjugated anti- mouse B220 (Southern Biotech, Birmingham, AL, USA), fluorescein confirming miR-125b expression using reverse transcriptase isothiocyanate- and phycoerythrin-conjugated anti-mouse Igk (BD Pharmingen, San Diego, CA, USA), phycoerythrin-conjugated anti-mouse quantitative PCR (Figure 1), transduced cells were co-cultured c-Kit and biotin-conjugated anti-mouse BP1 (both obtained from with OP9 cells for 6 days in the presence of IL-7. An equal number eBioscience, Inc., San Diego, CA, USA) and allophycocyanin-conjugated of cells from each group were taken (B70% of the streptavidin (BD Pharmingen). After staining, flow cytometry was B220 þ /EGFP þ population; Supplementary Figure S1) performed using a FACSCalibur or FACScan (both obtained from BD and viable B-cell progenitors were counted at days 3, 8 and 11 Biosciences, San Jose, CA, USA); data were analyzed with FlowJo Software after transduction. We found that miR-125b-overexpressing cells (Treestar, Ashland, OR, USA). Cell sorting was performed using a FACSAria II were constantly more abundant (Figures 2a and b) and at day 3 (BD Biosciences). the number of miR-125b-transduced pre-BI cells was 40% higher than control cells. This percentage continued to increase over Retroviral transduction time, and by day 11 miR-125b-overexpressing cells were three Retroviral supernatants were produced by the Phoenix retrovirus producer times as abundant as control cells (Figure 2b, Po0.01), suggesting line (Orbigen) using the Lipofectamine 2000 (Invitrogen) reagent for a constant growth advantage. We performed bromodeoxyuridine transfection, according to the manufacturers’ instructions. Then, cells were assays on pre-BI cells transduced with either miR-125b, miR-125b- treated with 10 mM Sodium Butyrate (Millipore, Billeria, MA, USA) and mut or the empty vector. At days 2, 4 and 6 after transduction, washed with phosphate-buffered saline before Pre-BI cell medium was bromodeoxyuridine incorporation into pre-BI cells was evaluated 1 added. Viral supernatants were harvested after 24 h of incubation at 32 C and we found that the rate of proliferating cells was higher in and filtered through a 0.45-mm filter. Pre-BI cells, cultured on OP9 stromal cells in the presence of IL-7, were transduced on the third day after miR-125b-overexpressing cells after day 4 (Figures 2c and d) isolation from the bone marrow. Retroviral supernatant was diluted (Po0.05 and Po0.01). In addition, we performed cell cycle 1:3, supplemented with polybrene to a final concentration of 3 mg/ml analysis of transduced cells and wild-type pre-BI cells (Sigma-Aldrich), then transduction of pre-BI cells, 70/Z3 and BaF/3 cells was (Supplementary Figure S2). As expected, the percentage of cells performed by spin infection at 3500 r.p.m., 32 1C for 120 min and overnight present in S phase was higher in pMIE-miR-125b-infected cells at incubation at 37 1C. day 4 compared with controls (Supplementary Figure S2). To check that the effects of miR-125b were not restricted to wild-type IL-7 starvation assay and B-cell differentiation analysis murine B-cell progenitors, we transfected murine (38B9, 18–81 Cells in IL-7 were washed then plated in Iscove’s Modified Dulbecco’s and 70Z/3) and human (Nalm6) pre-B-cell lines with pre-miR-125b Medium containing 5% FCS, 0.05 mM b-Mercaptoethanol and 2 mM or control microRNA. We quantified miR-125b expression in these L-Glutamine on an OP9 stroma cell layer. On day 6 cells were stained for cell lines before and after transfection (Figure 3a). At days 2, 3 surface BP1 and Igk light chain and analyzed by flow cytometry. and 4, we performed MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3- carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner Statistical analysis salt) proliferation tests and found that the miR-125b-expressing Results are shown as the mean±s.e.m. Statistical analysis was performed cells were more numerous in all four cell lines (Figure 3b). Taken using the Student’s t-test as provided by Microsoft Excel (Microsoft together, these results suggest that overexpression of miR-125b Corporation, Redmond, WA, USA) and the null hypothesis was rejected at enhances proliferation in wild-type pre-BI cells as well as in the 0.05 level (***Po0.0005, **Po0.005, *Po0.05). leukemic pre-B-cell lines, with these effects being more pro- Further information on Materials and Methods used is available in the nounced in non-leukemic contexts. Furthermore, the effects were Supplementary File. comparable between human and murine cell lines.

RESULTS miR-125b overexpression has anti-apoptotic effects independent miR-125b overexpression induces proliferation of wild-type of the p53 pathway in wild-type B-cell progenitors and in B-ALL progenitor B-cells and B-ALL cell lines cell lines To test the hypothesis that miR-125b could act as an oncogene by To test whether the increased cell number induced by miR-125b is increasing proliferation of progenitor B-cells, we first transduced the consequence of overproliferation or decreased apoptosis, we murine wild-type pre-BI cells with either an empty retroviral vector tested the influence of miR-125b on the activation of the effector (pMIE), or a vector expressing miR-125b-1 (pMIE-125b) or mutated caspases 3/7 and on the loss of plasma membrane asymmetry. miR-125b-1 (pMIE-125b-mut) in the seed region. We then counted Wild-type and pre-BI cells transduced with miR-125b, miR-125b- cells at different time points. Infected cells were distinguished by mut or empty vectors were sorted for B220 þ /EGFP þ cells, then the expression of enhanced green fluorescent protein. After seeded on irradiated OP9 cells and cultured in cell medium

& 2012 Macmillan Publishers Limited Leukemia (2012) 2224 – 2232 miR-125b targets ARID3a in B-ALL MP Puissegur et al 2226

Figure 2. miR-125b induces proliferation of pre-BI cells in vitro.(a) Microscopic images at day 4 of culture of pre-BI cells transduced with pMIE or pMIE-miR-125b in presence of IL-7. (b) Wild-type non-transduced pre-BI cells (NT) or pre-BI cells transduced with pMIE, pMIE-miR-125b or pMIE- miR-125b-mut vectors were cultured on an OP9 stromal layer in the presence of IL-7. At days 3, 8 and 11 of culture and following retroviral infection, cells were washed with phosphate-buffered saline, stained with propidium iodide to exclude dead cells and counted on a Malassez slide. Counting results were normalized each day by the number of NT cells. (c, d) Bromo-deoxyuridine incorporation assays were performed on non-transduced pre-BI cells, controls (pMIE, pMIE-miR-125b-mut) or cells expressing miR-125b (pMIE-miR-125b). (c) Representative FACS proﬁles of proliferative non-transduced and infected pre-BI cells (CD19 þ /EGFP þ /bromodeoxyuridine (BrdU) þ ) at day 4 of culture. (d) The mean percentage of proliferative non-transduced, pMIE, pMIE-miR-125b or pMIE-miR-125b-mut pre-BI cells at days 2, 4, 6 and 9 of culture from n ¼ 4 independent experiments are presented as means±s.d.

supplemented only with 2% FCS to induce stress conditions. lines. Western blot analysis revealed that neither p53 nor BAK1 Caspase activity was measured at days 2, 4 and 6 of culture, and expression levels significantly changed in each case (Figure 4c, we found that when compared with control cells at the same time Supplementary Figure S3), suggesting that neither BAK1 nor p53 point, significantly less caspase activity was found as time are targets of miR-125b in a murine pre-B cell context and that the progressed in miR-125b-positive cells (Figure 4a). To quantify the observed anti-apoptotic effect of this microRNA is caused by p53- rate of apoptosis, we performed annexin V staining on the same independent mechanisms. pre-BI cells. From day 4 we observed a significant decrease in apoptotic cell numbers in pre-BI cells overexpressing miR-125b (annexinV þ /7-aminoactinomycin D À ) when compared with con- Identification of an miR-125b target gene, ARID3A, which trol cells at the same time point (Figure 4b). Human and murine recapitulates the hyperproliferative effect of miR-125b in B-ALL cell lines were also assessed for caspase activation at days 1, 2, 3 cell lines and 4 after transfection using pre-miR-125b or pre-miR-neg control To elucidate the mechanism of miR-125b-mediated growth microRNA (Supplementary Figure S3). From day 2 onwards, advantage and protection from apoptosis in pre-BI cells, we caspase activity increased in both control and miR-125b-trans- tested the influence of miR-125b on the transcriptome in 70Z/3 fected cells but was always significantly lower in miR-125b-positive and 18–81 cells, using methodology described previously.12,13 As cells. Taken together, these results suggest that enforced expected, ectopic expression of miR-125b induced transcriptomic expression of miR-125b retards activation of the apoptotic alterations in 70Z/3 and 18–81 cells 48 hours after transfection. pathway and thus confers a survival advantage. Interestingly, using our web tool miRonTop (http:// Among the predicted targets of miR-125b in the apoptotic www.microarray.fr:8080/miRonTop/index),14 we found a specific process, one of the most interesting is p53. In humans and enrichment of miR-125b-predicted targets in the set of zebrafish, the 30 untranslated region of p53 mRNA has been shown downregulated transcripts. Whereas a moderate but significant to contain an active miR-125b-binding site,11 but in mice the enrichment was found with a basic complementary seed search situation is more complex as the miR-125b-binding site of p53 is (nucleotides 2–7 or nucleotides 2–8), a strong enrichment (460) not conserved. To evaluate whether the anti-apoptotic effect of was revealed using the validated microRNA targets database miR-125b resulted in p53 or BAK1 (a p53 effector) targeting, we miRTaRbase (Supplementary Figure S4). Overall, using several measured the protein levels of p53 and BAK1 3 days after additional prediction tools detailed in Table 1, an initial list of 28 transduction in miR-125b-positive or control pro-B and B-ALL cell putative candidates was produced. Among this list of predicted

Leukemia (2012) 2224 – 2232 & 2012 Macmillan Publishers Limited miR-125b targets ARID3a in B-ALL MP Puissegur et al 2227

Figure 3. miR-125b activates B-ALL cell line proliferation in vitro. (a) qRT-PCR analysis of miR-125b expression in B-ALL cells transfected with 10 nM of pre-miR-Neg or pre-miR-125b at day 2. The number on the y axis represents relative expression levels of mature miR-125b normalized to an internal control: sno202 for murine cell lines, 38B9, 18–81, 70Z/3 and RNU24 for the human cell line, Nalm6. (b) miR-125b induces B-ALL cell line proliferation in vitro. The effects of miR-125b on murine 38B9, 18–81, 70Z/3 cells and human Nalm6 cells were measured using the MTS assay. Cells were treated with 10 nM of pre-miR-Neg or pre-miR-125b from 1 to 5 days. Results correspond to three independent experiments performed in triplicate. *Po0.05 when compared with cells transfected with pre-miR-Neg. targets, which was established in 70Z/3 cells, a transcript encoding (Figure 5a, middle panel) and the reduction of Arid3a expression Arid3a was also found to be downregulated in 18–81 cells where at the protein level by western blot analysis in miR-125b- an approximate fourfold enrichment for miR-125b-validated expressing 70Z/3 cells (Figure 5a, right panel). In parallel we used targets was revealed using miRTaRbase (data not shown). a specific miR-125b antisense locked nucleic acid oligonucleotide ARID3a is a transcription factor also known as Bright in mouse to knockdown miR-125b in human leukemic REH cells, a cell line (B-cell regulator of immunoglobulin heavy-chain transcription). It derived from a patient with ETV6/RUNX1 leukemia that over- is a member of the ARID (A þ T-rich interaction domain) family of expresses miR-125b. After control of reduced miR-125b expression, proteins, which is required for both hematopoietic stem cell and we observed an increase both in transcription (Supplementary B-cell lineage development.15,16 To test whether Arid3a was Figure S5, left panel) and protein expression of ARID3a (Figure 5a, directly targeted by miR-125b, we fused the 30 untranslated region right panel). These data show definitively that ARID3a is a valid of Arid3a with and without three mutations in the miR-125b- target of miR-125b. binding site to the coding region of the luciferase reporter named We then investigated whether the hyperproliferative effects as pSI-Arid3a-mut and pSI-Arid3a, respectively. Only co- observed upon miR-125b expression in B-ALL cell lines might be transfection with pSI-Arid3a and a synthetic pre-miR-125b in the due to the downregulation of Arid3a expression. We transduced NIH/3T3 cells significantly decreased the luciferase activity when murine B-ALL 70Z/3 cells with a retrovirus expressing small hairpin compared with a pre-miR-negative control, strongly suggesting RNAs directed against Arid3a (pSh-Arid3a). After 2 days of that this gene is a target of miR-125b (Figure 5a, left panel). We puromycin selection, we confirmed the knockdown of Arid3a by then confirmed the repression of the Arid3a transcript by reverse reverse transcriptase quantitative PCR and western blot analysis transcriptase quantitative PCR in pMIE-125-transduced pre-BI cells (Figure 5a, right panel). Interestingly, from day 4 we observed that

& 2012 Macmillan Publishers Limited Leukemia (2012) 2224 – 2232 miR-125b targets ARID3a in B-ALL MP Puissegur et al 2228

Figure 4. miR-125b delays apoptosis activation in pre-BI cells without alteration of p53 and BAK1 expression levels. A caspase 3/7 assay (a) and phycoerythrin-AnnexinV staining (b) were carried out on pre-BI cells transduced under the same conditions as previously described and cultured in cell medium supplemented with 2% FCS. Results correspond to three independent experiments performed in triplicate. **Po0.005, *Po0.05. (c) Effect of miR-125b expression on endogenous p53 and BAK1 levels as determined by western blot on day 3 after transduction of pre-BI cells and on day 3 after transfection of B-ALL cell lines.

Table 1. List of the main miR-125b-predicted and miR-125b-validated targets downregulated following miR-125b-5p overexpression in the murine pre-BI cell line 70Z/3

Symbol Gene ID AveExpr logFC Seed 2–7 3’UTR Seed 2–8 3’UTR miRBase TargetScan Conserved miRTar Base v1

ABTB1 80283 12.25 À 1.66 X X X X X ARID3A 13496 9.81 À 1.56 X X X X X RHEBL1 69159 10.88 À 1.06 X X X — X RHOT2 214952 12.09 À 0.74 X X X X — PTPN18 19253 16.36 À 1.04 X X X X — BC003266 80284 13.30 À 0.58 X X X — — M6PR 17113 12.43 À 0.62 X X — X — MPHOSPH6 68533 12.40 À 0.88 X X X — — E2F2 242705 12.90 À 0.97 X X — X — C330006A16RIK 109299 12.00 À 0.58 X X — — — A230051G13RIK 216792 9.19 À 1.06 X X — — — WDR81 192652 11.25 À 0.59 X X — — — TMEM50B 77975 11.72 À 0.63 X X — — — 2310008H09RIK 66356 10.73 À 1.12 X X — — — HNRNPAB 15384 15.07 À 0.75 X X — — — DCP2 70640 8.22 À 1.03 X X — — — NAA40 70999 11.54 À 0.78 X X — — — DIDO1 23856 8.85 À 1.10 X X — — — D14ABB1E 218850 8.76 À 0.55 X — — — — ARL6IP6 65103 9.07 À 0.56 X — — — — WDR76 241627 7.03 À 0.77 X — — — — EXO1 26909 13.31 À 0.55 X — — — — BCL2L12 75736 11.56 À 0.72 X — — — — COQ2 71883 13.60 À 0.51 X — — — — MIER1 71148 7.94 À 0.70 X — — — — PCMTD1 319263 9.93 À 0.76 X — — — — SCML4 268297 7.54 À 1.02 X — — — — UBE2I 22196 10.71 À 0.78 X — — — — Abbreviation: UTR, untranslated region. The table lists the transcripts signiﬁcantly downregulated following miR-125b overexpression and those predicted to be miR-125b targets using the bioinformatics tool miRonTop (http://www.microarray.fr:8080/miRonTop/index) with different prediction algorithms: (i) an exact seed search 2–7 or 1–8 seed in the 30UTR; (ii) miRBase Target v5; (iii) TargetScan conserved miRTarBase v1 has been used to list the miR-125b-validated. Logarithm (base 2) of the average intensity (AveExpr) and logarithm (base 2) of the ratio of miR-125b/miR-Neg (logFC) are represented. Gene ID corresponds to the NCBI Gene IDs (http://www.ncbi.nlm.nih.gov/gene).

pSh-Arid3a-infected cells proliferate as well as pMIE-125b-trans- were validated using three different small hairpin RNAs against duced 70Z/3 cells and better than 70Z/3 cells transduced with a Arid3a to exclude the possibility of off-target effects. Moreover, we retroviral control (psh-Neg) (Figure 5b, left panel). These results tested whether repression of Arid3a would confer independence

Leukemia (2012) 2224 – 2232 & 2012 Macmillan Publishers Limited miR-125b targets ARID3a in B-ALL MP Puissegur et al 2229

Figure 5. miR-125b represses expression of Bright/Arid3a, which recapitulates the hyperproliferative effect of miR-125b in B-ALL cell lines. (a) NIH/3T3 cells were co-transfected with 10 nM pre-miR-Neg, pre-miR-125b and pSI-Arid3a constructs containing 30 untranslated region sequences with an miR-125b-binding site or a pSI-Arid3a-mut construct containing mutated seed miR-125b sequences in the 30 untranslated region of Arid3a (left panel). Arid3a expression was monitored by quantitative PCR on RNA from pMIE- and pMIE-miR-125b-transduced pre-B1 cells (middle panel) 2 days after transduction. Western blot analysis of Arid3a were performed on lysates extracted from pMIE-, pMIE-miR-125b-, pSh-Neg- or pSh-Arid3a-transduced 70Z/3 cells 2 days after transduction and after two supplementary days of selection by puromycin, which was necessary for small hairpin RNA expression and after knockdown of miR-125b in REH cells (right panel). (b) 70Z/3 cells and Ba/F3 cells were transduced with pMIE-, pMIE-miR-125b-, pSh-Neg-orpSh-Arid3a retroviral constructs. The MTS cell viability assay was performed from days 1 to 8 on transduced 70Z/3 cells (left panel). The transduced Ba/F3 cells were cultured in the absence of IL-3 for 24 h. After 24 h, IL-3 was added to the medium and cell proliferation was measured by MTS every 24 h, then normalized using cell proliferation of transduced Ba/F3 cells without deprivation of IL-3 (right panel). The graph shows an average of at least three different clones and three experiments. s.e. are indicated. **Po0.005, *Po0.05.

Figure 6. ARID3a is downregulated in B-ALL patients with an upregulation of miR-125b. Relative expression levels of ARID3a (a) and of four pluripotency-associated markers: OCT4, NANOG, SOX2 and KLF4 (b) were examined by quantitative PCR on RNA both from B-ALL patients with the recurrent IGH@ translocation, t(11;14)(q24;q32) that results in the upregulation of miR-125b and from B-ALL patients lacking this translocation, named as other B-ALL and previously described in Nixon et al.16 of IL-3 in a culture of murine Ba/F3 cells (Figure 5b, right panel). previously studied to discern whether or not they carry the We found that both pSh-Arid3a-transduced Ba/F3 cells and recurrent IGH@ translocation, t(11;14)(q24;q32), which results in an miR-125b-overexpressing Ba/F3 cells were highly resistant to a upregulation of miR-125b.9 Using RT-PCR we observed that ARID3a transient removal of IL-3. After reintroduction of IL-3 in the growth expression was lower in two patients with the translocation medium, only pMIE-125b and pSh-Arid3a-transduced Ba/F3 cells, compared to the B-ALL controls that do not overexpress miR-125b and not cells transduced with control vectors, resumed their (Figure 6a). normal growth. Thus, we have shown that Arid3a could be Recent studies support a model in which Bright/ARID3a acts as involved in the miR-125b-mediated growth advantage. a suppressor of lineage plasticity.15,17 Indeed, Bright/ARID3A inhibition causes increased developmental plasticity in mouse and human cells,15,17 notably by increasing the expression of ARID3a is repressed in B-ALL patients with the t(11;14)(q24;q32) pluripotency-associated factors such as Oct4, Sox2, Klf4 and translocation and induces expression of pluripotency-associated Nanog. To evaluate whether repression of ARID3a can induce the factors transcription of these markers, we tested their expression levels We next sought to examine whether ARID3a is downregulated in using reverse transcriptase quantitative PCR in B-ALL patients with primary leukemic blasts highly expressing miR-125b. We used or without the t(11;14) translocation. Interestingly, we noted a bone marrow samples from B-ALL patients that have been signiﬁcant upregulation of at least three pluripotency markers

& 2012 Macmillan Publishers Limited Leukemia (2012) 2224 – 2232 miR-125b targets ARID3a in B-ALL MP Puissegur et al 2230 (NANOG, SOX2 and KLF4) in patient P2 (t(11;14) þ ) (Figure 6b). Overexpression of these markers was not observed in patient P1; however, the bone marrow sample of patient P1 was removed after the relapse causing his death. Therefore, the patient had received chemotherapy treatment that could have altered the expression of these markers. To confirm a role for ARID3a and miR- 125b on the upregulation of pluripotency markers, we quantified and compared their expression in pMIE-125b-orpSh-Arid3a- transduced 70Z/3 cells. All four pluripotency markers were found to be overexpressed in cells whose expression of Arid3a had been repressed using either microRNA (pMIE-125b) or by small interfering RNA (pSh-Arid3a) (Supplementary Figure S5). In parallel, we knocked down miR-125b in human REH cells and observed a good correlation between the increased expression of ARID3a and the reduction of OCT4, SOX2, KLF4 and NANOG transcripts (Supplementary Figure S5). Thus, our data show that, in B-ALL, miR-125b upregulation is associated with an ARID3a downregulation and an increase in the expression of pluripotency-associated markers. Figure 7. Total repression of Arid3a blocks in vitro differentiation of murine B-cell progenitors. Wild-type pre-BI cells, pMIE, pMIE-miR- 125b-, pMIE-miR-125b-mut-, pMIE-PAX-ELN-, pSh-Neg- and pSh-Arid3a- miR-125b overexpression does not block differentiation of B-cell infected pre-BI cells were co-cultured with OP9 stroma cells in the progenitors in vitro presence or absence of IL-7 then stained at day 6 of culture using The second distinct event occurring in B-ALL transformation is the anti-BP1 and anti-Igk antibodies to differentiate B-cell progenitors (BP1 þ /Igk À ) from pre-BII cells (BP1 À /Igk À ) and immature B-cells blockage of differentiation at the pre-B-cell stage, therefore we À þ tested whether overexpression of miR-125b and repression of (BP1 /Igk ). Histogram represents the percentage of each cell type. Data are taken from n ¼ 3 independent experiments at day 6 of Arid3a altered B-cell differentiation. Murine wild-type pre-BI culture and are presented as means±s.d. cells were transduced with either retroviral constructs: pMIE, pMIE-miR-125b, pMIE-miR-125b-mut, pSh-Neg or pSh-Arid3a; or with a vector expressing PAX5-ELN (pMIE-PAX5-ELN) as a positive control. PAX5-ELN is a chimeric gene isolated from B-ALL patients miR-125b overexpression in B-ALL cells could cause a reduction in in our laboratory by Bousquet et al.,18 and is associated with a Arid3a levels and subsequent blockage of B-cell differentiation, blockage in differentiation at the pre-B-cell stage. Transduced cells thus contributing toward B-ALL transformation. were co-cultured for 6 days with OP9 cells in the presence or absence of IL-7 (IL-7 starvation assay19). Usually IL-7 withdrawal causes pre-BI cells to stop proliferating, start differentiating and DISCUSSION eventually undergo apoptosis in vitro.19 We used expression levels Recent reports of cases of B-ALL presenting with an insertion of of BP1 and surface Igk to determine which B-cell progenitors were miR-125b-1 into the IGH locus8–10 prompted us to analyze the present under each condition at day 6 of culture. BP1 is a 140-kDa effect of miR-125b-1 overexpression on the development of B-cell type II transmembrane protein with aminopeptidase A activity and progenitors. Here we have shown that miR-125b overexpression is expressed from the early pro-B to the late pre-B stages, with could contribute to leukemogenesis by conferring a proliferative expression diminishing after that. In contrast, surface Igk is advantage and anti-apoptotic capacity that is independent of p53 detected only during the later stages of differentiation. Cells were and directly targets ARID3a, a transcriptional activator of the stained and EGFP-positive living cells were analyzed by flow heavy-chain immunoglobulin gene.20 B-ALL transformation cytometry. In the presence of IL-7, 60% of B-cells expressed BP1 requires a blockage in differentiation as well as a clonal and only 3% expressed Igk (Figure 7, Supplementary Figure S6), expansion of malignant cells with a significant growth reflecting the persistence of B-cell progenitors at the pre-BI stage advantage. Using the pre-B-cell differentiation model developed (BP1 þ /Igk À ). In the presence of IL-7, BP1 and Igk expression by A Rolink,19 we found that enforced expression of miR-125b in levels were comparable between non-transduced cells and cells murine pre-BI cells did not have a significant effect on B-cell transduced by pMIE, pMIE-miR-125b, pMIE-miR-125b-MUT and differentiation. However, in our experiments only a moderate pMIE-PAX5-ELN (Figure 7, Supplementary Figure S6). However, expression (fivefold over mean) of miR-125b occurred in pre-BI upon IL-7 withdrawal around 20% of cells expressed BP1 and cells and led to a moderate repression of Arid3a. We suppose that B35% of this population were expressing Igk (BP1 À /Igk þ ), this repression was not strong enough to induce the blockage in indicating that differentiation of the B-cell progenitors had taken B-cell differentiation as pre-BI cells with a total repression of place. In contrast, 45% of control cells transduced with the Arid3a (pSh-Arid3a-transduced pre-BI cells) did not acquire Igk,a positive control pMIE-PAX5-ELN þ expressed BP1 in the absence of marker for the late stages of B-cell differentiation. In support of IL-7. This reflects the ability of the chimeric PAX5-ELN protein to this, it has previously been shown that B-ALL patients with the block B-cell differentiation (Po0.01). Loss of BP1 and acquaintance t(11;14)(q24;q32) translocation have a 30- to 600-fold higher of Igk expression were comparable between non-transduced, expression of miR-125b compared with B-ALL cases without this control and miR-125b-expressing cells (Figure 7, Supplementary translocation.9 Figure S6), suggesting that this microRNA had no significant Here, we have shown that miR-125b overexpression leads not impact on normal B-cell differentiation. However, pre-BI cells in only to higher proliferation but also to remarkable inhibition of which Arid3a was knocked down were blocked at the pre-BII stage apoptosis. This was the case in wild-type progenitor B-cells as well of differentiation upon IL-7 deprivation, as they lacked the BP1 as in murine and human pre-B-cell lines. A recent publication on surface marker and did not express Igk (Figure 7, Supplementary miR-125b-2 in ETV6/RUNX1 B-ALL showed a survival advantage Figure S6). conferred by overexpression of miR-125b.7 These findings are in These results show that repression of Arid3a can block B-cell line with our results; however, our report is the first to examine the differentiation in vitro and suggests a mechanism by which effect of miR-125b in a wild-type pre-B-cell model, which is much

Leukemia (2012) 2224 – 2232 & 2012 Macmillan Publishers Limited miR-125b targets ARID3a in B-ALL MP Puissegur et al 2231 closer to in vivo conditions as wild-type pre-B-cells are non- repressed. Thus, in B-ALL overexpressing miR-125b several lines of modified, non-transformed cells and are directly isolated from the evidence suggest that this microRNA targets ARID3a/Bright and bone marrow of healthy mice. Furthermore, the pre-B-cell model maintains the progenitor characteristics of pre-BI cells by is a dynamic system that allows us to test the effects of this controlling the expression of pluripotency-associated factors. microRNA on B-cell growth. Interestingly, the observed effects Furthermore, the repression of ARID3a could lead to inhibition regarding proliferation and apoptosis were more pronounced in of differentiation of pre-BI cells by preventing the transcription of wild-type pre-BI cells compared with leukemic cell lines, which immunoglobulin heavy-chain genes. A recent study reported that underscores the value of this model. in pediatric ALL cases, overexpression of miR-125b is associated In this study, we also tested whether the observed anti- with resistance to vincristine, particularly in patients carrying apoptotic effect of miR-125b was because of direct targeting of the the TEL-AML1 translocation.37 Therefore, it would be interesting pro-apoptotic proteins p53 or BAK1. In fact, both of these proteins to investigate the mechanisms of cellular resistance to chemo- are predicted targets of miR-125b from bioinformatics-based therapy in malignancies with an up-regulation of miR-125b. prediction tools, and a recent publication further suggests that the whole p53 network is a conserved target of miR-125b.21 In this study, seven genes were identified, including upstream regulators CONFLICT OF INTEREST of p53 and downstream effectors such as BAK1, that all contain The authors declare no conflict of interest. putative binding sites for miR-125b.21 Although these are not all conserved across species, each vertebrate species has a binding site for miR-125b in at least one of these genes, underlining the ACKNOWLEDGEMENTS 21 importance of the p53 network as a target of miR-125b. In We acknowledge the excellent support of the Nice-Sophia Antipolis Functional addition, p53 and BAK1 have already been experimentally proven Genomics Platform for microarray processing. This work was supported by the as targets of miR-125b in humans.4,11 However, when we tested Fondation pour la Recherche Me´dicale, a Carl-Duisberg scholarship from the Bayer whether p53 and BAK1 expression levels were affected by miR- Science and Education Foundation, the Institut National du Cancer, the Association 125b expression in progenitor B-cells, we found no change in their pour la Recherche sur le Cancer (ARC), the CITTIL program and the Institut protein levels. Although p53 is not a direct target of miR-125b in Universitaire de France. mice, it can be indirectly affected in its expression by upstream regulators as shown in mouse fibroblasts.11 However, in a pre-B- AUTHOR CONTRIBUTIONS cell context, we showed here that p53 and BAK1 are not targeted by miR-125b, hence the anti-apoptotic effects we observed are MPP, RE, CC and EC performed the experiments, BM, KL, CB and MB helped probably related to other mechanisms. Our results are reinforced design experiments and interpret results, FNK provided patients’ samples. MPP, by data by Gefen et al.7 on miR-125b-2 in ETV6/RUNX1 leukemia, RE and PB designed the study and wrote the manuscript. where the conferred survival advantage is also p53-independent. MicroRNAs are highly versatile molecules. This is partially explained by the fact that a single microRNA can have hundreds REFERENCES of targets, which are differentially expressed depending on the 1 Cobaleda C, Sanchez-Garcia I. B-cell acute lymphoblastic leukaemia: towards cell type.22 Thus, miR-125b has been implicated in mechanisms as understanding its cellular origin. Bioessays 2009; 31: 600–609. diverse as metabolism,23 the innate immune response,24 neuronal 2 Calin GA, Dumitru CD, Shimizu M, Bichi R, Zupo S, Noch E et al. Frequent deletions 25,11 26,27 5,9,10,28 and down-regulation of micro-RNA genes miR15 and miR16 at 13q14 in chronic differentiation, hematopoiesis and cancer. In lymphocytic leukemia. Proc Natl Acad Sci USA 2002; 99: 15524–15529. cancer, miR-125b is reported to have different effects in different 3 Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 2004; 29 tumors. For example, tumor suppressor activity in breast cancer 116: 281–297. 30 and ovarian carcinoma, as well as oncogenic properties in 4 Shi XB, Xue L, Yang J, Ma AH, Zhao J, Xu M et al. An androgen-regulated miRNA pancreatic cancer,31 prostate cancer,4 myelodysplastic syndrome suppresses Bak1 expression and induces androgen-independent growth of and acute myeloid leukemia,5 megakaryoblastic leukemia32 and prostate cancer cells. Proc Natl Acad Sci USA 2007; 104: 19983–19988. B-ALL,7,33 have all been related to this microRNA. miR-125b can 5 Bousquet M, Quelen C, Rosati R, Mansat-De Mas V, La Starza R, Bastard C et al. even have opposite effects on the same process as it augments Myeloid cell differentiation arrest by miR-125b-1 in myelodysplastic syndrome proliferation in prostate cancer cells4 but inhibits cell cycling in and acute myeloid leukemia with the t(2;11)(p21;q23) translocation. J Exp Med 16 2008; 205: 2499–2506. hepatocellular carcinoma, suggesting that miR-125b might act as 6 Zhang H, Luo XQ, Zhang P, Huang LB, Zheng YS, Wu J et al. MicroRNA patterns an oncogene and/or a tumor suppressor, depending on the associated with clinical prognostic parameters and CNS relapse prediction in cellular context. To identify pertinent miR-125b targets in a pre-B- pediatric acute leukemia. PLoS One 2009; 4: e7826. cell context, we performed gene expression profiling of a leukemic 7 Gefen N, Binder V, Zaliova M, Linka Y, Morrow M, Novosel A et al. Hsa-mir-125b-2 B-cell line using this microRNA. We showed that the ARID3a/Bright is highly expressed in childhood ETV6/RUNX1 (TEL/AML1) leukemias and confers gene is a direct target of miR-125b at both mRNA and protein survival advantage to growth inhibitory signals independent of p53. Leukemia levels, in both murine B-cell progenitor cells as well as in B-ALL cell 2010; 24: 89–96. lines. Moreover, in the same cellular context this repression 8 Sonoki T, Iwanaga E, Mitsuya H, Asou N. Insertion of microRNA-125b-1, a human mimics the proliferative and anti-apoptotic effects induced by homologue of lin-4, into a rearranged immunoglobulin heavy chain gene locus in a patient with precursor B-cell acute lymphoblastic leukemia. Leukemia 2005; 19: miR-125b. 2009–2010. Our study emphasizes the fact that miR-125b can act as an 9 Chapiro E, Russell LJ, Struski S, Cave´ H, Radford-Weiss I, Valle VD et al. Anew oncogene in progenitor B-cells by targeting ARID3a/Bright, a recurrent translocation t(11;14)(q24;q32) involving IGH@ and miR-125b-1 in B-cell transcription factor implicated in the regulation of expression of progenitor acute lymphoblastic leukemia. Leukemia 2010; 24: 1362–1364. the immunoglobulin heavy chain20 and in cellular plasticity.15,16 10 Tassano E, Acquila M, Tavella E, Micalizzi C, Panarello C, Morerio C et al. MicroRNA- Interestingly, in the mouse Arid3a/Bright, expression is tightly 125b-1 and BLID upregulation resulting from a novel IGH translocation in child- regulated and is restricted to B-cells but it is not expressed in all hood B-Cell precursor acute lymphoblastic leukemia. Genes Chromosomes Cancer human B-lymphocyte subpopulations. Indeed, in human cells 2010; 49: 682–687. Bright is expressed in early B-lymphocyte precursors (pro-B and 11 Le MT, Teh C, Shyh-Chang N, Xie H, Zhou B, Korzh V et al. MicroRNA-125b is a novel negative regulator of p53. Genes Dev 2009; 23: 862–876. pre-B cells) and in germinal center cells,34,35 places where miR- 36 12 Pottier N, Maurin T, Chevalier B, Puisse´gur MP, Lebrigand K, Robbe-Sermesant K et 125b overexpression has already been described. al. Identification of keratinocyte growth factor as a target of microRNA-155 in lung Finally, in B-ALL patients with the t(11;14)(q24;q32) transloca- fibroblasts: implication in epithelial-mesenchymal interactions. PLoS One 2009; 4: tion, we have shown that the expression of ARID3a is dramatically e6718.

& 2012 Macmillan Publishers Limited Leukemia (2012) 2224 – 2232 miR-125b targets ARID3a in B-ALL MP Puissegur et al 2232 13 Puisse´gur MP, Mazure NM, Bertero T, Pradelli L, Grosso S, Robbe-Sermesant K et al. 25 Schickel R, Boyerinas B, Park SM, Peter ME. MicroRNAs: key players in the miR-210 is overexpressed in late stages of lung cancer and mediates mitochon- immune system, differentiation, tumorigenesis and cell death. Oncogene 2008; 27: drial alterations associated with modulation of HIF-1 activity. Cell Death Differ 5959–5974. 2011; 18: 465–478. 26 Chaudhuri AA, So AY, Mehta A, Minisandram A, Sinha N, Jonsson VD et al. 14 Le Brigand K, Robbe-Sermesant K, Mari B, Barbry P. MiRonTop: mining microRNAs Oncomir miR-125b regulates hematopoiesis by targeting the gene Lin28A. Proc targets across large scale gene expression studies. Bioinformatics 2010; 26: 3131–3132. Natl Acad Sci USA 2012; 109: 4233–4238. 15 Webb CF, Bryant J, Popowski M, Allred L, Kim D, Harriss J et al. The ARID family 27 O’Connell RM, Chaudhuri AA, Rao DS, Gibson WS, Balazs AB, Baltimore D. transcription factor bright is required for both hematopoietic stem cell and B MicroRNAs enriched in hematopoietic stem cells differentially regulate long-term lineage development. Mol Cell Biol 2011; 31: 1041–1053. hematopoietic output. Proc Natl Acad Sci USA 2010; 107: 14235–14240. 16 Nixon JC, Ferrell S, Miner C, Oldham AL, Hochgeschwender U, Webb CF. 28 Bousquet M, Harris MH, Zhou B, Lodish HF. MicroRNA miR-125b causes leukemia. Transgenic mice expressing dominant-negative bright exhibit defects in B1 B Proc Natl Acad Sci USA 2010; 107: 21558–21563. cells. J Immunol 2008; 181: 6913–6922. 29 Scott GK, Goga A, Bhaumik D, Berger CE, Sullivan CS, Benz CC. Coordinate sup- 17 An G, Miner CA, Nixon JC, Kincade PW, Bryant J, Tucker PW et al. Loss of pression of ERBB2 and ERBB3 by enforced expression of micro-RNA miR-125a or Bright/ARID3a function promotes developmental plasticity. Stem Cells 2010; 28: miR-125b. J Biol Chem 2007; 282: 1479–1486. 1560–1567. 30 Nam EJ, Yoon H, Kim SW, Kim H, Kim YT, Kim JH et al. MicroRNA expression 18 Bousquet M, Broccardo C, Quelen C, Meggetto F, Kuhlein E, Delsol G et al. A novel profiles in serous ovarian carcinoma. Clin Cancer Res 2008; 14: 2690–2695. PAX5-ELN fusion protein identified in B-cell acute lymphoblastic leukemia acts as 31 Bloomston M, Frankel WL, Petrocca F, Volinia S, Alder H, Hagan JP et al. MicroRNA a dominant negative on wild-type PAX5. Blood 2007; 109: 3417–3423. expression patterns to differentiate pancreatic adenocarcinoma from normal 19 Fazio G, Palmi C, Rolink A, Biondi A, Cazzaniga G. PAX5/TEL acts as a transcrip- pancreas and chronic pancreatitis. JAMA 2007; 297: 1901–1908. tional repressor causing down-modulation of CD19, enhances migration 32 Klusmann JH, Li Z, Bo¨hmer K, Kremens B, Vormoor J, Dworzak M et al. miR-125b-2 to CXCL12, and confers survival advantage in pre-BI cells. Cancer Res 2008; 68: is a potential oncomiR on human chromosome 21 in megakaryoblastic leukemia. 181–189. Genes Dev 2010; 24: 478–490. 20 Herrscher RF, Kaplan MH, Lelsz DL, Das C, Scheuermann R, Tucker PW. The 33 Enomoto Y, Kitaura J, Hatakeyama K, Watanuki J, Akasaka T, Kato N et al. TEm/miR- immunoglobulin heavy-chain matrix-associating regions are bound by Bright: 125b transgenic mice develop lethal B-cell malignancies. Leukemia 2011; 25: a B cell-specific trans-activator that describes a new DNA-binding protein family. 1849–1856. Genes Dev 1995; 9: 3067–3082. 34 Webb CF, Smith EA, Medina KL, Allred L, Kim D, Harriss J et al. Expression of 21 Sinha AU, Kaimal V, Chen J, Jegga AG. Dissecting microregulation of a master bright at two distinct stages of B lymphocyte development. J Immunol 1998; 160: regulatory network. BMC Genomics 2008; 9: 88. 4747–4754. 22 Filipowicz W, Bhattacharyya SN, Sonenberg N. Mechanisms of post-transcriptional 35 Nixon JC, Rajaiya JB, Ayers N, Evetts S, Webb CF. The transcription factor, Bright, is regulation by microRNAs: are the answers in sight? Nat Rev Genet 2008; 9: not expressed in all human B lymphocyte subpopulations. Cell Immunol 2004; 102–114. 228:42–53. 23 Tili E, Michaille JJ, Calin GA. Expression and function of micro-RNAs in immune 36 Gururajan M, Haga CL, Das S, Leu CM, Hodson D, Josson S et al. MicroRNA cells during normal or disease state. Int J Med Sci 2008; 5: 73–79. 125b inhibition of B cell differentiation in germinal centers. Int Immunol 2010; 22: 24 Tili E, Michaille JJ, Cimino A, Costinean S, Dumitru CD, Adair B et al. Modulation of 583–592. miR-155 and miR-125b levels following lipopolysaccharide/TNF-alpha stimulation 37 Schotte D, De Menezes RX, Moqadam FA, Khankahdani LM, Lange-Turenhout E, and their possible roles in regulating the response to endotoxin shock. J Immunol Chen C et al. MicroRNA characterize genetic diversity and drug resistance in 2007; 179: 5082–5089. pediatric acute lymphoblastic leukemia. Haematologica 2011; 96: 703–711.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

Leukemia (2012) 2224 – 2232 & 2012 Macmillan Publishers Limited