miR-125b modulates megakaryocyte maturation by targeting the cell-cycle inhibitor p19INK4D

A better understanding of the mechanisms involved in megakaryocyte maturation will facilitate the generation of platelets in vitro and their clinical applications. A microRNA, miR-125b, has been suggested to have important roles in the self-renewal of megakaryocyte-erythroid progenitors and in platelet generation. However, miR-125b is also critical for hematopoietic stem cell self-renewal. Thus, the function of miR-125b and the complex signaling pathways regulating megakaryopoiesis remain to be elucidated. In this study, an attentive examination of the endogenous expression of miR-125b during megakaryocyte differentiation was performed. Accordingly, the differentiation of hematopoietic stem cells requires the downregulation of miR-125b, whereas megakaryocyte determination and maturation synchronize with miR-125b accumulation. The overexpression of miR-125b improves megakaryocytic differentiation of K562 and UT-7 cells. Furthermore, stage-specific overexpression of miR-125b in primary cells demonstrates that miR-125b mediates an enhancement of megakaryocytic differentiation after megakaryocyte determination, the stage at which megakaryocytes are negative for the expression of the hematopoietic progenitor marker CD34. The identification of miR-125b targets during megakaryopoiesis was focused on negative regulators of cell cycle because the transition of the G1/S phase has been associated with megakaryocyte polyploidization. Real-time PCR, western blot and luciferase reporter assay reveal that p19INK4D is a direct target of miR-125b. P19INK4D knockdown using small interfering RNA (siRNA) in megakaryocyte-induced K562 cells, UT-7 cells and CD61+ promegakaryocytes results in S-phase progression and increased polyploidy, as well as improved megakaryocyte differentiation, similarly to the effects of miR-125b overexpression. P19INK4D overexpression reverses these effects, as indicated by reduced expression of megakaryocyte markers, G1-phase arrest and polyploidy decrease. P19INK4D knockdown in miR-125b downregulated cells or p19INK4D overexpression in miR-125b upregulated cells rescued the effect of miR-125b. Taken together, these findings suggest that miR-125b expression positively regulates megakaryocyte development since the initial phases of megakaryocyte determination, and p19INK4D is one of the key mediators of miR-125b activity during the onset of megakaryocyte polyploidization.

Thrombocytopenia, the deficiency of platelets (PLTs) in the blood, threatens millions of people, including patients undergoing high-dose chemotherapy, and subjects affected by aplastic anemia or hepatitis virus-related cirrhosis. The cells responsible of PLT production are the megakaryocytes (MK). Polyploidization is an important step during MK maturation and PLT generation. To understand the mechanisms underlying MK maturation will facilitate ex vivo PLT manufacture for therapeutic applications and clinical treatments of thrombocytopenia.
MicroRNAs (miRNAs) are small non-coding RNA molecules that regulate gene expression primarily by inhibiting the translation of target mRNAs through direct binding of specific sites in the 3′-untranslated region (3′-UTR). 1 It is now commonly accepted that miRNAs has essential roles in hematopoiesis, including embryonic stem cell differentiation, erythropoiesis, granulocytopoiesis/monocytopoiesis, lymphopoiesis and megakaryocytopoiesis. 2 During MK determination and maturation, miR-155 blocks megakaryocytic differentiation by targeting Ets-1 and Meis1 transcription factors, miR-150 drives MK-erythroid precursors toward the megakaryocytic fate via the inhibition of the target transcription factor c-Myb and miR-34a enhances MK differentiation in hematopoietic stem cells (HSCs) through the repression of c-Myb and MEK1 expression. 3 Recently, Klusmann et al. 4 reported that the overexpression of the onco-miR miR-125b-2 could induce the proliferation and differentiation of MK and MK-erythroid precursors. The same authors also revealed that miR-125b enhanced the polyploidization and PLT output of MKs. 5 Further, miR-125b-2 was found to have a direct correlation with the pathogenesis of acute megakaryoblastic leukemia in Down syndrome patients. Therefore, it is crucial to uncover the correlation between miR-125b and megakaryocytopoiesis. Although the study by Klusmann et al. 4 suggests that Dicer and ST18 were direct targets of miR-125b in megakaryocytopoiesis, both of these factors are common to several cellular processes. 6 Considering the significant effect of miR-125b on cell proliferation and endomitosis of MK progenitors and MKs, it is reasonable to speculate that miR-125b may regulate the expression of cellcycle inhibitors. In this study, we systematically identified the role of miR-125b in the differentiation of MKs and one of the target genes that mediate its function, the cell-cycle inhibitor p19 INK4D .

Results
Endogenous expression of miR-125b increases during MK differentiation and PLT production. To gain insight into the role of miR-125b in MK differentiation, we investigated the endogenous expression of miR-125b in HSC-MK differentiation. First, we examined the total cellular expression of miR-125b in differentiating megakaryoid cells. RNA was collected every 3 days, starting from day 0 over 15 days of induction. To evaluate megakaryocytic maturation and enrichment, we analyzed the expression of the megakaryocytic surface markers CD41 (GPIIb) and CD61 (GPIIIa/β3). The elevated expression of these MK biomarkers (CD41 and CD61) in the cultures is shown in Figure 1a. miR-125b expression decreased during the first 6 days of in vitro differentiation, and then slightly increased after day 6 of culture. The expression of miR-125b was markedly elevated in PLTs isolated from cord blood (CB) (Figure 1b) (4200-fold) compared with undifferentiated CD34 + hematopoietic cells (Figure 1d, left panel). Although miR-125b expression in primary cells exhibited a certain degree of variability among the individual donors, it progressively and markedly increased in PLTs in all of the samples analyzed.
We then attempted to separate mature and immature MKs based on the different fluorescence intensity of the MK markers. CB cells cultured for 15 days were chosen for FACS sorting. Three populations were separated according to the expression levels of CD41 and CD61 (Figure 1c). These populations are designated as: CD41 hi CD61 hi (CD41 ++ CD61 ++ , P2), CD41 med CD61 med (CD41 + CD61 + , P3) and CD41 med CD61 low (CD61 + , P4). Sorted MKs at distinct stages of development, starting from CD34 + hematopoietic cells to cells cultured for 15 days, also present morphological differences. From HSC to P2, the cell size increased and the nucleus of the MKs became larger and lobulated. miR-125b expression levels were low in the early stages of differentiation, and progressively increased until the terminal phases of maturation. Compared with CD34 + HSCs, we also found an enrichment (10-fold) of miR-125b in CD41 hi CD61 hi late-stage MKs, suggesting that miRNA-125b might be involved in the common process of MK terminal differentiation and PLT generation (Figure 1d, right panel).
K562 and UT-7 cells can be induced to differentiate into cells with megakaryoblastic characteristics by phorbol myristate acetate (PMA) treatment. 7,8 K562-PMA and UT-7-PMA models were used to validate the results obtained from human CD34 + CB hematopoietic cells. We compared miR-125b expression in undifferentiated versus PMA-treated cells. After PMA treatment, K562 and UT-7 cells stopped proliferating, became larger and polyploid and expressed the MK markers CD41 and CD61 (Figures 2a-c). This phenotype suggests that PMA treatment facilitates the production of MKs. When comparing miRNA expression in untreated cells and in cells cultured for 3 days with PMA, an upregulation of miR-125b by 6.5-fold was observed after PMA treatment in K562 cells (Figure 2d, top panel). Similarly, megakaryocytic differentiation of UT-7 cells resulted in the enrichment (10-fold) of miR-125b (Figure 2d, bottom panel). Thus, MK differentiation of HSCs, K562 cells and UT-7 cells was associated with an increased expression of miR-125b. These observations suggest that miR-125b might be positively associated with megakaryocytic differentiation.
Effect of miR-125b overexpression or downregulation on MK differentiation. Endogenous expression of miR-125b is low in two human leukemia cell lines, K562 and UT-7. Therefore, we overexpressed miR-125b in K562 and UT-7 cells to determine whether the upregulation of miR-125b observed during MK differentiation has a role in controlling MK maturation. Successful overexpression of mature miR-125b in both cell lines is shown in Figure 3a. Subsequently, the cells were cultured in medium containing 1 nM PMA to induce megakaryotic differentiation and maturation. Overexpression of miR-125b promoted the expression of MK integrins (Figure 3c). Another indicator of MK differentiation is the expression of genes important for MK lineage, which regulate lineage-specific differentiation of MKs and characterize the maturation of human MKs (megakaryocytic genes). Similarly to what was observed by flow cytometry, higher expression of CD61 and other important megakaryocytic genes was also detected by quantitative RT-PCR (qPCR) in pcDNA3.1-pri-miR-125b cells when compared with control cells, which is consistent with previous observations (Figure 3b).
We further determined whether miR-125b also contributes to MK differentiation of primary human hematopoietic stem and progenitor cells. For this purpose, miR-125b expression was modified in human CB mononuclear cells (MNCs). The cells were grown in megakaryocytic differentiation medium and transfected on day 1 with miR-125b mimics or non-target negative control (NC) mimics. Transfection of miR-125b mimics increased miR-125b levels of is~30-fold compared with NC cells (Figure 4a). Then, we evaluated the effect of miR-125b overexpression on MK markers and megakaryocytic genes. After 6 days of culture, miR-125b-transfected MNCs showed a higher CD61 + (49.4% versus 43.1%) and CD41 + / CD61 + percentage (20.6% versus 17.9%) than NC cells (Figure 4b). The mRNA expression of MK glycoproteins (CD41 and CD61) increased 42-fold. Moreover, qPCR results showed that the expression of MK transcription factors (GABPA, GATA-1, GATA-2, FIL1 and RUNX) and thrombopoietin (TPO) receptor (C-MPL) increased accordingly ( Figure 4c).
Subsequently, MNCs were transfected with miR-125b inhibitor mimics or non-target NC inhibitors. By using specific inhibitors, miR-125b expression was successfully downregulated over fivefold in MNCs ( Figure 4d). As expected, the miR-125b inhibitors retarded MK differentiation. The percentage of CD61 + and CD41 + /CD61 + populations in MNCs transiently transfected with miR-125b inhibitors was 36.3% APC 15day 6    During megakaryopoiesis, the expression of CD61 showed the most marked and progressive increase over time, while surface expression of CD34 decreased. 10 Using wellestablished surface markers, we purified three populations of cells. We refer to them as: HSCs CD34 + CD61 − , neonatal MK progenitors (CFU-MKs) CD34 + CD61 + and mature MK progenitors (pro-MKs) CD34 − CD61 + . Magnetic cell sorting (MACS) technology was used to isolate the three populations from human CB MNCs before or after MK differentiation in vitro. After separation, we stained the cells with antibodies against CD34 and CD61, and examined the endogenous expression of miR-125b (Figures 5a-c). Subsequently, miR-125b expression was modified in HSCs, CFU-MKs and pro-MKs (Figures 5d, g, j, m, p and s), followed by 5 days MKspecific induction and detection.
For HSCs and CFU-MKs, gain or loss of function of miR-125b did not result in a significant alteration of the percentage of CD61 + and CD41 + populations and cellular morphology after the induction of differentiation (Figures 5d-o and Supplementary Figures S1A and B).
Studies with promegakryocytes demonstrated that both early (CD41 and CD61) and late (CD42b) megakaryocytic surface markers could be detected by flow cytometry after induction. And the overexpression of miR-125b promoted the expression of all these MK differentiation markers (Figure 5q and Supplementary Figure S1C). When comparing MK maturity by cellular morphology, pro-MKs with miR-125b mimics show more mature MK phenotypes, the huger and higher polyploid (8N or more), than NC mimics (Figure 5r). Moreover, miR-125b inhibitor significantly retarded MK differentiation as assessed by comparing cellular morphology and expression of the megakaryocytic surface markers (Figures 5t-u). Concomitantly, we evaluated the number of PLT released in culture media. From the same number of seeded pro-MKs, we obtained more PLTs in miR-125btransfected pro-MKs than the NC cells and less in miR-125b decreased expression than with NC inhibitor (Supplementary Figures S1F-G).
To confirm the function of miR-125b at distinct stages of MK development, a CFU-MK colony formation assay was also performed to quantitate the megakaryocytic potential of the progenitor cells. As we known, neonatal MK progenitors are hyperproliferative and generate more MKs per colony, although the colony cells are smaller and less polyploid compared with adult progenitors that generate more PLTs per MK. 11 As a consequence, large MK colonies arise from more primitive MK progenitors, whereas smaller MK colonies are produced by more mature MK progenitors. Therefore, we subdivided the colonies based on their size: small (3-20 cells per colony), medium (21-49 cells per colony) and large (⩾50 cells per colony). We found that in HSCs, forced miR-125b expression resulted in more CFU-GEMM and large CFU-MK units, but had no effect on medium and small CFU-MK units (Figures 5v-x), suggesting a uniform effect of miR-125b on all the myeloid lineages at this stage. Sorted CFU-MKs and pro-MKs could not derive myeloid colonies except for megakaryocytic colonies (data not shown). For these cells, miR-125b overexpression resulted in a statistically significant increase in total MK colony-forming units, especially in small CFU-MK units, but showed no effect on large CFU-MK units. In contrast, loss of miR-125b function showed a marked decrease in megakaryocytic colony formation (Figures  5w-z). In conclusion, these gain-and loss-of-function experiments strongly suggest that miR-125b promotes MK differentiation at a late stage of maturation.
The G1 cell-cycle inhibitor p19 INK4D is a target of miR-125b during megakaryopoiesis. In silico prediction of miRNA targets suggests that thousands of mRNAs could be regulated by miR-125b at the transcriptional or posttranscriptional level. Few studies have noted that miR-125b has different functions in different cellular contexts. 12 In megakaryocytopoiesis, a previous report 5 and our study both demonstrated the important role that miR-125b has in MK endomitosis, which results in the polyploidy of MKs. Accordingly, we narrowed our analysis of potential miR-125b targets in megakaryopoiesis to cell-cycle regulators, especially to the INK4 family, whose members have been suggested to have a pivotal role in G1 cell-cycle suppression and MK endomitosis. 13 By consulting relevant literatures and performing integrative bioinformatic analysis, we identified potential miR-125b binding sites in the 3′-UTRs of p16 INK4A , p15 INK4B and p19 INK4D (Figure 6a). We had demonstrated the upregulation of miR-125b was required for the maturation of MKs, as a consequence, its putative targets should be downregulated during this process. During human MK differentiation, an upregulation of p15 INK4B

3' a g u g u u C A A U C C C A G A G U C C C u 5' hsa-miR-125b
| | | | | : | | | | | | 269 5' g g g g u c G U U U G C U U U U C A G G G u 3' p15 INK4B

3' a g u g U U C A A U C C C A G A -G U C C C u 5'
hsa-miR-125b CD34 + hematopoietic cells during MK differentiation and PLT production. We found that the expression of p19 INK4D first increased and then decreased. In particular, p19 INK4D mRNA also decreased in PLTs and in CD41 hi CD61 hi late-stage MKs (Figure 6b). Similarly, following PMA treatment, mRNA expression of p19 INK4D decreased during MK differentiation (Figure 6c). Further, a negative correlation between the expression of miR-125b and p19 INK4D was shown. Next, we sought to verify the regulation of p19 INK4D by miR-125b.

| | | | | | | | | | | | | 310 5' a g u g A A G U C A U G A C C U C C A G G G a 3'
Overexpressing miR-125b in CD34 + hematopoietic cells decreased the mRNA expression levels of p19 INK4D . A remarkable increase of p19 INK4D mRNA was observed in cells transfected with miR-125b inhibitors compared with cells transfected with NC inhibitors (Figure 6d). Protein and mRNA expression levels of p19 INK4D in pcDNA3.1-pri-miR-125b/ K562 cells were considerably lower compared with that in pcDNA3.1 cells (Figure 6e). Therefore, p19 INK4D is one of the putative target genes of miR-125b, which has an identical seed sequence complementary to the binding site on the 3′-UTR of p19 INK4D . Using a luciferase reporter assay, we demonstrated that p19 INK4D is a direct target of miR-125b. miR-125b overexpression significantly decreased the luciferase activity of the reporter containing the wild-type 3′-UTR of p19 INK4D , when compared with the mutant form and the controls (Figure 6f).
Overexpression or knockdown of p19 INK4D inhibits or promotes MK differentiation, respectively. P19 INK4D is an important regulator of cell-cycle progression. Previous reports showed that p19 INK4D expression is correlated with endomitotic arrest and MK maturation and regulated by AML-1. 17 Taken together with the results that p19 INK4D mRNA level decreased not only during MK differentiation but also in PLTs, we speculated that this molecule is involved in the arrest of MK terminal differentiation. To confirm our hypothesis, two human leukemia cell lines, K562 and UT-7 cells and CB-derived CD61 + pro-MKs were transfected with siRNAs against p19 INK4D , and examined the effect of RNAi-mediated knockdown of p19 INK4D (Figure 7a). After siRNA transfection and megakaryocytic induction, sip19 INK4D -transfected cells showed higher CD41 + /CD61 + rate compared with control cells (Figure 7b and Supplementary Figures S2A and B). Moreover, the transcriptional levels of megakaryocytic genes also increased upon p19 INK4D knockdown (Figure 7c). Cell-cycle variation is a key feature of endomitosis. We next analyzed cell proliferation and polyploid level after p19 INK4D knockdown. In agreement with the effect observed following miR-125b overexpression, cell proliferation was significantly reduced and the cell size and ploidy level was increased along with the decreased expression of p19 INK4D during MK differentiation (Figures 7d and e). Furthermore, ploidy analysis showed that p19 INK4D reduction increased the percentage of 4N and 8N cells (Figure 7f) together with the G1-phase abatement (Figure 7g).
We next examined whether miR-125b inhibition and subsequently p19 INK4D upregulation as well as MK differentiation blockage could be antagonized by direct p19 INK4D knocking down (Figure 7h). Both in K562 cells and pro-MKs, p19 INK4D siRNA treatment partially rescued the elevation of p19 INK4D mRNA, resulting from miR-125b inhibitor transfection (Figure 7i). The antidifferentiation effects of miR-125b inhibitor in megakaryocytic surface markers and cellular morphology were, in part, abrogated by the concomitant p19 INK4D downmodulation (Figures 7j and k and Supplementary Figures S2C and D).
To confirm the role of p19 INK4D in MK differentiation and polypoidization, the effect of p19 INK4D upregulation on MK development was detected. For p19 INK4D overexpression, the expression plasmid pcDNA3.1-p19 INK4D -puromycin was constructed and transfected into K562 and UT-7 cells. MNCs were infected with HA-tagged p19 IK4D -expressing retrovirus. The upregulation of p19 INK4D was confirmed at both the mRNA and protein levels (Figures 8a, b and Supplementary Figure S3A). P19 INK4D overexpression significantly delayed MK differentiation, as assessed by the CD41 + /CD61 + and MK polyploidization rate. Especially, 62.5% of the p19 INK4D -transfected MNCs (HA + ) were CD41/61 + on average versus 72.1% for the control virus-infected cells, which implies that p19 INK4D overexpression is capable of inhibiting MK differentiation (Figures 8c, d and Supplementary Figure S3C). The cell-cycle assay demonstrated that p19 INK4D inhibits cellular proliferation by increasing the G1 compartment (Figure 8e). When p19 INK4D expression was rescued by pcDNA3.1-p19 INK4D transfection in miR-125b-overexpressing K562 cells (Figures 8f and g), the MK promoting effects of miR-125b were, in part, abrogated by the complementary p19 INK4D overexpression (Figures 8h, i and Supplementary Figure S3D). Thus, we showed by integrative analysis that miR-125b may exert its effect on MK differentiation by repressing p19 INK4D expression.

Discussion
MiR-125b is currently considered as a key molecule in the physiology and pathology of stem cell self-renewal and tumorigenesis. Ooi et al. 18 and O'Connell et al. 19 elucidated that miR-125b is a critical regulator of HSC self-renewal. It was also demonstrated by Klusmann et al. 4 that miR-125b promotes MK self-renewal. Because miR-125b has been suggested to participate in both HSC self-renewal and MK differentiation, we tried to clarify the stage-specific effect of this molecule. Enhanced miR-125b expression promotes HSC self-renewal, as suggested by increased CFU content. Differentiation toward the myeloid lineage and MKs requires the downregulation of miR-125b over a specific period of time ( Figure 1c). However, as soon as the determination of MKs has been achieved, miR-125b downregulation stops, and this miRNA keeps on accumulating until the final stages of MK differentiation, when PLTs are produced. It is reasonable to speculate that the intracellular microenvironment has an important role in the distinct function of miR-125b. In our study, the phase-specific function of miR-125b was clarified by stage-specific miRNA gene modification. Based on our findings, the specificity of miR-125b function to megakaryopoiesis should start at the late stages of MK development, in which miR-125b synergizes with MK-specific regulators (transcriptional factors, phosphorylated signal molecules and on on) to further accelerate the maturation of MKs. These cooperating factors remain to be determined in further studies.
Cell-cycle regulators have been reported to have an important role in hematopoiesis. 20 We demonstrated that the  . Dunnett's test T3 were used for statistical analysis. Cells were obtained from four different donors and each sample was tested with three independent experiments for data presentation. The results are presented as mean ± S.D. Individual comparisons between two groups were performed using Student's paired t-test. Bonferroni's multiple comparison test for multiple comparisons was applied. *Po0.05, **Po0.01 and ***Po0.001 (scale bars: 50 μm) p19 INK4D overexpression resulted in megakaryocytic differentiation blockage, as well as G1-phase arrest. As a conclusion, miR-125b improves the maturation and polyploidization of MKs by inhibiting p19 INK4D expression. However, the function of the accumulation of miR-125b in PLTremains to be determined. It is possible that the suppression of cyclindependent kinases by miR-125b mediates the final process of MK maturation. Taken together, our findings contribute to elucidate the specific effect of miR-125b in MK development. miR-125b acts as a positive regulator of megakaryopoiesis, and supports the MK fate during all of the phase of HSC, MEP and MK differentiation. The effect on MK differentiation and polyploidization is, at least partially, mediated by the downregulation of p19 INK4D . The application of miR-125b or small molecules that regulate its downstream signaling pathway might partially solve the shortage of PLT generation ex vivo, and meet the urgent requirement of PLT in the clinic. This study may contribute to the development of therapeutic strategies for PLT diseases.

Materials and Methods
Purification of human CB-derived CD34 + cells. Human CB samples were obtained from umbilical cord after the delivery of normal pregnancies with patient's informed consent. Research has been carried out in accordance with the approval of the ethics committee. CB samples were processed within 4 h. MNCs were isolated by Ficoll-Hypaque density gradient centrifugation (1.077 g/l; TBDscience, Tianjin, China). Then, CB-CD34 + cells were purified from the MNCs by magnetic force following the manufacturer's instructions (MACS; MiltenyiBiotec, Bergisch Glabach, Germany).
Cell culture and reagents. Human erythroleukemia K562 cells were cultured in RPMI-1640 medium supplemented with 10% fetal calf serum. Human megakaryoleukemia UT-7 cells were cultured in RPMI-1640 medium supplemented with 10% fetal calf serum and 1 U/ml Epo. For MK differentiation, cells were treated with 1 nM PMA (Sigma) for 3 days.
For microscopy imaging, cells were spun on microscopic slides and stained with Wright-Giemsa solutions. NIKON microscope and NIS-Elements F software (Nikon Corporation,Tokyo, Japan) were used to capture picture and measure cell diameter.
Cell proliferation assay. Cell proliferation was analyzed with Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan). As described in the manufacturer's instructions, 100 μl of K562 or UT-7 cell suspension (2 × 10 4 cells per well) was dispensed in a 96-well plate. The cells were transfected by siRNA 1 day before the assay initiation. Subsequently, the plates were incubated with or without PMA (1 μM) at different times (0, 24, 48 and 72 h). Ten microliters of the CCK8 solution was added to the well of the plate. After 2 h incubation at 37°C, the absorbance at 450 nm was measured by a microplate reader.
Western blots were probed with anti-p19 INK4D , anti-β-actin or anti-HA antibodies purchased from Santa Cruz (Santa Cruz, CA, USA). β-Actin was used as a loading control.
Flow cytometric analysis for cell cycle and polyploidy. Cells were collected and washed in PBS, followed by permeabilization with cold 70% methanol for preserving at − 20°C. Propidium iodide (50 μg/ml) with RNAse were added to stain DNA in the cells for 30 min, and analyzed by BD FACS Calibur. Ploidy mode was used in polyploid analysis miRNA mimics and inhibitors transfection. Cells were seeded in 6well plates and transfected by nucleofection with 4 μg per well plasmid DNA, 1.6 μg per well miRNA mimics (Genepharma, Shanghai, China), or non-target NC mimics according to the manufacturer's protocol.
Quantitative RT-PCR. As described in the manufacturer's instructions, total RNA was isolated with TRIzol (Invitrogen, Carlsbad, CA, USA). Then, RNA (0.8 μg) was reverse transcribed using a PrimeScript RT-PCR Kit (Takara Bio, Shiga, Japan). Quantitative RT-PCR was performed in triplicate using the SYBR Green Master Mix (Toyobo, Osaka, Japan) in a Real-Time PCR Machine (Bio-Rad, Hercules, CA, USA). All gene expression levels were normalized to the housekeeping gene GAPDH. Analysis of mature miRNA expression relative to U6 snRNA was performed using All-in-One qPCR Mix (GeneCopoeia, Rockville, MD, USA).The specific primers are listed in Supplementary Table 1.
Dual-luciferase reporter assay. The cloned sequences and putative binding sites of miR-125b on p19 INK4D 3′-UTR are described in Figure 6. The sequences were cloned into the pGL3 control vector (Promega, Madison, WI, USA). For the luciferase reporter assay, stable miR-125b overexpressing or not K562 cells were transfected by 0.8 μg plasmid (PGL3, wild-type p19 INK4D -PGL3 or mutant p19 INK4D -PGL3), together with the pRL Renilla luciferase vector (Promega) as an internal control. Luciferase activities were assayed 48 h post-transfection according to the manufacturer's protocol.
Retrovirus production, concentration and cell infection. Retrovirus was produced with plat-A cells, which are stably expressing virus packing proteins. The day before transfection, 5 × 10 6 plat-A cells were plated in poly-L-lysine-coated 10 cm dishes. Twenty micrograms of retroviral vectors, MSCV-N-FLAG-HApuromycin (MSCV-N) or MSCV-N-FLAG-HA-p19 INK4D -puromycin (MSCV-p19), were used for transfection. Retroviral supernatant was concentrated by ultracentrifugation (26 000 r.p.m., 90 min). Titers of retrovirus were tested on 293T cells and HA and p19 INK4D expression was analyzed by western blot 3 days after infection (Supplementary Figure S3C). A total of 3 × 10 6 MNCs were infected with high titer p19 IK4D -expressing retrovirus overnight in 1 ml megakaryocytic differentiation medium with 8 μg/ml polybrene. The medium was changed one day after transduction. HA and cellular surface marker was analyzed by flow cytometric analysis 15 days after transduction.
Statistics. Selective statistic analysis is indicated in each experiment.

Conflict of Interest
The authors declare no conflict of interest.