MiR‐486 promotes proliferation and suppresses apoptosis in myeloid cells by targeting Cebpa in vitro

Abstract The monocytic MDSC (M‐MDSC) is one of the major types of MDSCs, which play important roles in suppression of antitumor immunity. However, the mechanisms underlying how M‐MDSCs so heavily accumulate in patients with cancer are still poorly understood. The purpose of this study was to identify miRNAs that regulate the proliferation and differentiation of M‐MDSCs. Microarray analysis was performed to identify differentially expressed miRNAs between tumor‐induced M‐MDSCs (TM‐MDSCs) and their counterparts from tumor‐free mice. The miRNAs and their target genes that regulate the proliferation and differentiation of myeloid cells were predicted by bioinformatics analysis and validated by RT‐qPCR. Luciferase reporter assays were used to analyze the relationships between miRNAs and target genes. Overexpression of candidate miRNAs and target genes in myeloid cells was conducted to verify their functions in cell proliferation, differentiation, and apoptosis. Our data showed that miR‐486 was overexpressed in TM‐MDSCs. Cebpa was predicted to be one of the target genes of miR‐486 that regulates the proliferation of myeloid cells. Expression of Cebpa was inversely correlated with miR‐486 in TM‐MDSCs, and we found that overexpression of miR‐486 suppressed the expression of Cebpa in both 293T cells determined by luciferase reporter assays and in myeloid cells determined by RT‐qPCR. Overexpression of miR‐486 promoted proliferation and suppressed apoptosis in myeloid cells, as opposed to overexpression of Cebpa, which promoted the opposing phenotype. Overexpression of either miR‐486 or Cebpa inhibited differentiation of myeloid cells. This study indicates that miR‐486 promotes proliferation and suppresses apoptosis in myeloid cells by targeting Cebpa in vitro, suggesting that miR‐486 and Cebpa might be involved in the expansion of TM‐MDSCs in tumor‐bearing mice.


| BACKGROUND
Tumor-induced myeloid-derived suppressor cells (MDSCs) are involved in tumor angiogenesis and evasion of host immunity through T cell suppression. It has been reported that the number of MDSCs in the peripheral blood correlates with prognosis and tumor burden. The phenotype of MDSCs in mice is Gr1 + CD11b + . 1 Murine MDSCs can be divided into two subsets based on the expression of Ly6C and Ly6G. Monocytic MDSCs (M-MDSCs), whose phenotype is CD11b + Ly6G − Ly6C hi/+ , have a monocyte-like morphology and express high levels of inducible nitric oxide synthase (iNOS) and arginase 1 (Arg1). 2,3 Granulocytic MDSCs (G-MDSCs), whose phenotype is CD11b + Ly6G + Ly6C low/+ , not only have a granulocyte-like morphology, preferentially producing reactive oxygen species (ROS) and expressing Arg1, but also possess a polymorphonuclear (PMN) morphology, explaining why G-MDSCs are also abbreviated to PMN-MDSCs. 2,3 Although G-MDSCs account for the majority of MDSCs, M-MDSCs have been found to be no less immunosuppressive than G-MDSCs on a per cell basis. 2,4,5 A third subset of MDSCs, namely, more immature, or early, MDSCs (eMDSCs) was recently defined, and the eMDSC phenotype is Lin − (CD3/14/15/19/56) HLA-DR − CD33 + in humans and has yet to be determined in mouse. 6 Substantial evidence from preclinical and clinical studies has revealed that the expansion of MDSCs in the tumor microenvironment is mediated by a combination of factors, including cytokines and growth factors. [7][8][9] Although micro ribonucleic acids (miRNAs), which are small noncoding RNAs that posttranscriptionally regulate the production of proteins, account for only 1%-2% of all genes in mammals, each miRNA is predicted to regulate hundreds of targets. Therefore, the majority of protein coding genes are considered to be under the regulation of miRNAs. Accumulative evidence has shown that miRNAs regulate many cell processes, such as proliferation, differentiation and survival. 10 Moreover, it was reported that miRNAs also modulate hematopoietic lineage differentiation, 11 a typical example of which includes myeloid-specific miR-223 inhibition of progenitor proliferation, granulocyte differentiation and activation. 12 There have been several studies validating the hypothesis that MDSCs originating from myeloid cells might also be regulated by miRNAs. [13][14][15][16] Studies have shown that both miR-17-5p and miR-20a decrease the suppressive potential of G-MDSCs, while having no effect on M-MDSCs. 16 IL-6-stimulated CD11b + CD14 + HLA-DR − M-MDSCs are associated with progression and poor prognosis in squamous cell carcinoma of the esophagus. 17 It was recently reported that ipilimumab treatment decreases M-MDSCs and increases CD8 effector memory T cells in long-term survivors with advanced melanoma, indicating that the frequency of M-MDSCs can predict treatment outcome. 18 As M-MDSCs possess special biological characteristics from G-MDSCs, the miRNA expression profile of tumor-induced M-MDSCs (TM-MDSCs) should theoretically be different from that of tumor-induced G-MDSCs. Therefore, it is necessary to distinguish M-MDSCs from G-MDSCs when exploring the miRNAs that regulate their functions. In this study, we performed miRNA microarray analysis on both M-MDSCs from tumor-bearing mice (TM-MDSCs) and their counterparts (CD11b + Ly6G − Ly6C hi cells) from tumor-free mice to identify differentially expressed miR-NAs between TM-MDSCs and their counterparts. Target genes of the identified miRNAs were predicted, among which genes that regulate proliferation and differentiation of myeloid cells were found by bioinformatics analyses. The functions of candidate target genes and associated miRNAs were further validated experimentally.

| Animals and establishment of tumorbearing mouse models
Female C57BL/6 mice, 4-6 weeks of age, were purchased from the department of laboratory animal science of Fudan University (Shanghai, China) and were housed under specific pathogen-free (SPF) conditions. Tumor-bearing mouse models were created by injecting 1 × 10 6 LLC cells or 1 × 10 5 B16 cells subcutaneously into the flank. When tumor size reached 15-20 mm in diameter, within 3-4 weeks after injection, mice were euthanized with sodium pentobarbital. All animal experiments were conducted in accordance with the protocol for animal use and handling. This study was approved by the Animal Welfare and Ethics Committee of the Department of Laboratory Animal Science, Fudan University (permit no. 20150295A010).

| RNA extraction and miRNA microarray assay
M-MDSCs and their counterparts were stored in Trizol LS (Catalog number: 1596-026; Invitrogen, Carlsbad, California, USA). Total RNA, including miRNAs, was isolated and purified using a miRNeasy Mini Kit (Catalog number: 217004; Qiagen, Hilden, Germany) and RNasefree DNase I (Catalog number: 79254; Qiagen) following the manufacturer's instructions. Qualified RNA samples were measured using the miRNA array using GeneChip miRNA 3.0 Array (Catalog number: 902017; Affymetrix, Santa Clara, California, USA). Labeling and hybridization were conducted following the manufacturer's protocols. A GeneChip Scanner (GeneChip2 Scanner 30007G; Affymetrix) was used to scan signals on the slides. Feature extraction was performed using Affymetrix GeneChip Command Console software (AGCC). Raw data were processed using Expression Console Software (Affymetrix). First, raw data was normalized using a quantile algorithm. Probes with at least 100.0 percent of samples in any 1 condition out of 2 conditions that were flagged as "Detected" were chosen for further analysis. Differentially expressed miRNAs were then identified by fold change as well as P-value calculation using the t test. Threshold for up-and down-regulated genes was set as a fold change ≥2.0 and a P value ≤ 0.05. MicroRNA expression profiles of granulocytic MDSCs have been reported in our previous publication. 19

| Bioinformatics analysis
To identify miRNAs and their target genes that regulate proliferation and differentiation in M-MDSCs, we predicted the target genes of differentially expressed miRNAs between TM-MDSCs and their counterparts screened by microarray assay using miRwalk online software (http://www.umm.uniheidelberg.de/apps/zmf/mirwalk). We also selected genes that regulate the proliferation and differentiation of myeloid cells using Ingenuity Pathway Analysis (IPA) online software (http://www.ingenuity.com/products/ipa). We integrated genes identified by both miRwalk and IPA software, and only overlapping genes were considered as candidates. Thus, corresponding miRNAs were considered candidate miRNAs that could regulate proliferation and differentiation of tumor-induced M-MDSCs and myeloid cells.

| Real-time quantitative PCR
Total RNA was isolated from cells using TRIzol ® (Catalog number: 1596-026; Invitrogen) according to the manufacturer's protocol. RNA yield was determined using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Waltham, Massachusetts, USA), and integrity was evaluated using agarose gel electrophoresis stained with ethidium bromide. Quantification was performed with a two-step reaction process: reverse transcription (RT) and PCR. RT reactions were performed in a GeneAmp ® PCR System 9700 (Applied Biosystems, Foster City, California, USA) for 60 minutes at 37°C, followed by heat inactivation of RT for 5 minutes at 95°C. PCR reactions were incubated in a 384-well optical plate (Roche, Basel, Swiss) at 95°C for 10 minutes, followed by 40 cycles of 95°C for 10 seconds, 60°C for 30 seconds. Complementary DNA was synthesized from 1 μg total RNA using a miScriptII Reverse Transcriptase Mix (Catalog number: 218161; Qiagen). Real-time quantitative polymerase chain reaction (RT-qPCR) was conducted using the LightCycler 480 SYBR Green I Master kit (Roche) protocol on a LightCycler 480 II RT-PCR platform (Roche). Amplification of U6 small RNA (for mature miR-486) and GAPDH mRNA (for CCAAT/enhancer-binding protein-alpha [Cebpa]) was performed with each experimental sample as an endogenous control. All reactions were run in triplicate. Relative expression levels were calculated using the comparative 2 −ΔΔCt method. Primer sequences were designed and synthesized by Generay Biotech (Generay, Shanghai, PRC) based on mRNA sequences obtained from the miRBase database (for miR-486) and NCBI database (for Cebpa). The primer sequence of miR-486 was 5′-tcctgtactgagctgccccgag-3′, and the reverse primer of miR-486 was provided by Qiagen (Catalog number: 218161; Qiagen), which is proprietary. The forward primers for Cebpa and GAPDH were 5′-tgagtgaggctctcattctt-3′ and 5′-atcactgccacccagaag-3′, respectively. The reverse primers for Cebpa and GAPDH were 5′-acatacacccttggacaacta-3′ and 5′-cagggatgatgttctgggca-3′, respectively.

| Transfection of myeloid cells with lentivirus
Total nucleated myeloid cells, that had been treated with GM-CSF and IL-6 for 4 days, were cultured in 6-well plates at 3 × 10 5 /mL per well for 24 hours and transfected the following day without GM-CSF and IL-6. Lentivectors were added to infect myeloid cells for 48 hours and these cells were subsequently analyzed for proliferation, differentiation and apoptosis.

| Western blotting analysis
Cells were lysed using RIPA lysis buffer (Catalog number: BYL40825; JRDUN Biotech, Shanghai, China) containing both protease and phosphatase inhibitors. Cell lysates for western blot analysis were denatured and subjected to SDS-PAGE. According to the molecular weight of Cebpa, which is 43 kDa, we chose a 10% gel. The semi-dry method was used to transfer membranes, and 30 μg per well of protein was utilized. Proteins were transferred to nitrocellulose filter membranes (Catalog number: HATF00010; Millipore, USA), and membranes were blocked with 5% skim milk powder overnight at 4°C. Membranes were then incubated with primary antibodies against Cebpa (Catalog number: ab40764; Abcam, Cambridge, UK) and GAPDH (Catalog number: 5174; CST, Boston, Massachusetts, USA), which were diluted 500 times and 1500 times, respectively, in PBS containing 5% skim milk powder, according to their manual instructions and were incubated with membranes at room temperature for 2 hours. Membranes were then hybridized with secondary HRP-conjugated antibodies (Catalog number: A0208; Beyotime Biotechnology), which were diluted 1000 times according to manual instruction in TBST at 37°C for 1 hour. Protein-Ab complexes were detected using enhanced chemiluminescence (ECL; Catalog number: WBKLS0100; Millipore). The luminescence signal was visualized on a chemiluminescent imaging system (Model: Tanon 5200; Tanon Science & Technology Company Limited, Shanghai, China).

| Statistical analysis
Results are presented as the mean ± SD. Data were analyzed using GraphPad Prism 6 software for Windows (GraphPad Software, Inc, La Jolla, CA, USA). Statistical analyses were performed using Stata/SE 10.1 software (Statacorp, College Station, Texas, USA). A two-tailed Student's t test was used to compare data between two groups, and one-way ANOVA was used to compare data between three groups with the Scheff multiple-comparison test. For miRNA microarray assay, fold changes of 2 or more and a P-value < 0.05 were considered statistically significant. For all other tests, P < 0.05 was considered statistically significant.

| Differentially expressed miRNAs between TM-MDSCs and their counterparts
M-MDSCs from the spleens of LLC-bearing mice (TM-MDSCs) and their counterparts (CD11b + Ly6G − Ly6c hi cells from spleens of tumor-free mice) were purified using flow cytometry. The gating strategy used to isolate M-MDSCs is shown in Figure S1 and in our previous paper. 19 MiRNA expression profiles were generated using Affymetrix GeneChip miRNA 3.0 Array. The results of the two-way hierarchical clustering of genes and samples are displayed as a heat map diagram ( Figure 1A), which was generated with the MutiExperiment Viewer (MEV) software. Five differentially expressed miRNAs were detected between TM-MDSCs and their counterparts in the microarray assay. MiR-486, miR-3107, miR-192, and miR-99b were expressed at higher levels in TM-MDSCs, and miR-3096b was expressed at a lower level in TM-MDSCs than their counterparts. Figure 1A also shows another 17 microRNAs that were not differentially expressed between TM-MDSCs and their counterparts but that were overexpressed in tumor-induced G-MDSCs compared with their counterparts (CD11b + Ly6G + Ly6c low/+ cells) from tumor-free mice (Fold changes ≥ 2.0 and P-value < 0.05 found by microarray). 19 The five differentially expressed miR-NAs between TM-MDSCs and their counterparts as detected by microarray are also displayed by scatter plot ( Figure 1B). To explore these miRNAs and their target genes that regulate the proliferation and differentiation of TM-MDSCs, we predicted target genes for the differentially expressed miRNAs between TM-MDSCs and their counterparts using miRwalk online software. Since MDSCs are derived from myeloid cells, we picked genes that regulate the proliferation (108 genes) and differentiation (346 genes) of myeloid cells using IPA online software. We integrated genes identified by both miRwalk and IPA software, and only overlapping genes were considered candidates. Cebpa was predicted to be the only gene that could regulate the proliferation of TM-MDSCs and myeloid cells, and the corresponding miRNA was miR-486.

| Overexpression of miR-486 promotes proliferation, and suppresses apoptosis and differentiation in myeloid cells
Myeloid cells from mice, that included all myeloid cells, with the exception of red blood cells, were transfected with vehicle lentivirus or miR-486-expressing lentivirus for 48 hours, or were left without treatment (blank control). Cells were examined by flow cytometry. The results showed that overexpression of miR-486 promoted the proliferation of myeloid cells (Figure 2A) and suppressed the apoptosis of myeloid cells ( Figure 2B). The expressions levels of CD11c, CD80, CD86, F4/80, and CD11b were suppressed by miR-486. Both CD11b + Ly6G + Ly6C low cells (phenotype of G-MDSCs) and CD11b + Ly6G − Ly6C hi cells (phenotype of M-MDSCs) were comparable between the three groups. The data indicate that overexpression of miR-486 inhibits myeloid differentiation ( Figure 2C-H).

| Expression of miR-486 is inversely correlated with the potential target gene Cebpa in M-MDSCs
To explore the relationship between miR-486 and Cebpa in M-MDSCs, we measured miR-486 and Cebpa expression in TM-MDSCs from both LLC-bearing mice and B16-bearing mice. The results showed that the expression of miR-486 in TM-MDSCs from both LLC-bearing mice and B16 melanoma-bearing mice was higher than that in their counterparts ( Figure 3A), which also validated our microarray data. Additionally, the expression of Cebpa was lower in TM-MDSCs from both LLC-bearing mice and B16 melanoma-bearing mice than in their counterparts ( Figure 3B).

| Overexpression of miR-486 inhibits expression of Cebpa
To assess whether miR-486 directly regulates the expression of Cebpa, a fragment of the 3′-UTR of Cebpa mRNA, containing the putative miR-486-binding sequence (wild type) or a mutated binding site (mutant), was cloned into a luciferase reporter construct and cotransfected with a control renilla luciferase reporter plasmid into 293T cells along with control or miR-486 mimic. Renilla luciferase activity served as an internal control. We found that the miR-486 mimic, but not the control miRNA, significantly reduced renilla luciferase activity in the wild-type 3′-UTR Cebpa groups; however, neither the miR-486 mimic nor the control miRNA significantly reduced renilla luciferase activity in the mutant 3′-UTR Cebpa groups, suggesting that miR-486 regulates gene expression directly at the Cebpa 3′-UTR ( Figure 3C). We next investigated the role of miR-486 in the regulation of Cebpa in myeloid cells. Murine myeloid cells were isolated and cultured in RPMI-1640, and then, the cells were infected with miR-486-expressing lentivector, vehicle lentivirus, or left without treatment (blank) for 48 hours. The transfection efficiency of the murine myeloid cells infected with vehicle lentivirus or miR-486-expressing lentivirus was observed by fluorescence microscopy ( Figure 3D). The expression of Cebpa in these cells was measured by both Western-blot and RT-qPCR. The results indicated that the overexpression of miR-486 inhibits the expression of Cebpa in myeloid cells ( Figure 3E,F). Original uncropped and unprocessed western blot images are provided in Figure S3.

| Overexpression of Cebpa induces apoptosis and suppresses proliferation and differentiation of myeloid cells
Murine myeloid cells were transfected with vehicle lentivirus or Cebpa-expressing lentivirus for 48 hours, and overexpression of Cebpa was subsequently confirmed by RT-qPCR ( Figure 4A). Overexpression of Cebpa suppressed proliferation ( Figure 4B) and promoted apoptosis in myeloid cells ( Figure 4C). Furthermore, overexpression of Cebpa reduced the expression levels of CD11c, CD80, CD86, and F4/80 but had no effect on the percentage of CD11b + Ly6G + Ly6C low cells and CD11b + Ly6G − Ly6C hi cells compared with the vehicle group, indicating that overexpression of Cebpa inhibits myeloid cell differentiation ( Figure 4D-I).

| DISCUSSION
MDSCs are a heterogeneous population of activated immature myeloid cells that are characterized by a morphological mixture of granulocytic and monocytic cells but lack B16 melanoma-bearing mice was higher than in control groups (their counterparts from spleens of tumor-free mice) as determined by RT-qPCR. B, Expression of Cebpa, which was predicted as one target of miR-486 by bioinformatics, was lower in TM-MDSCs from both LLC-bearing mice and B16 melanoma-bearing mice than in control groups. C, Interaction between miR-486 and Cebpa was detected using the Promega luciferase reporter assay, showing that the miR-486 mimic, but not control miRNA, significantly reduced renilla luciferase activity in wide type 3′-UTR Cebpa groups; however, neither the miR-486 mimic, nor control miRNA, significantly reduced renilla luciferase activity in mutated 3′-UTR Cebpa groups. D, Transfection efficiency in murine myeloid cells infected with vehicle lentivirus or miR-486-expressing lentivirus observed by a fluorescence microscope. Original magnification ×100. E, Murine myeloid cells were infected with vehicle lentivirus or miR-486-expressing lentivirus for 48 h. Expression of Cebpa in these cells was measured by western blot. The results also indicated that miR-486 reduced the expression of Cebpa in myeloid cells. F, Overexpression of miR-486 decreased the expression of Cebpa in myeloid cells as shown by RT-qPCR. Three tumor-bearing mice and five tumor-free mice were used for each test in A and B. All data are presented as the mean ± SD and repeated three times. Error bars represent SD. Each red scatter plot overlaid onto the solid bar graphs indicates one technical repeat. For A and B, a two-tailed Student's t test was used to compare data between two groups. For C and F, the P-values above the horizontal lines without vertical bars represent the results of one-way ANOVA for three groups, and other P-values represent the results of the Scheff multiple-comparison test when the P-values were lower than 0.05 in the one-way ANOVA analyses Since two main characteristics of MDSCs are marked accumulation and immaturity, understanding the mechanisms that regulate the proliferation and differentiation of MDSCs is important for targeting these cells. Preliminary data from previous studies has shown that miRNAs seem to have the potential to increase MDSC accumulation in both human cancers and tumor-bearing mice. 7,13 As M-MDSCs have some biological characteristics that are different from those of G-MDSCs, the miRNAs that regulate proliferation and differentiation of M-MDSCs may be different from those of G-MDSCs. To screen candidate miRNAs that regulate the proliferation and differentiation of TM-MDSCs, we performed miRNA microarray analysis to identify differentially expressed miRNAs between TM-MDSCs and their counterparts (CD11b + Ly6G − Ly6c hi cells from tumor-free mice). Five miRNAs were identified, and of these, miR-486 was overexpressed in TM-MDSCs compared with their counterparts. MicroRNA expression profiles of G-MDSCs compared with their counterparts from tumor-free mice have been reported in our previous publication. 19 According to the criteria that fold changes should be no less than 2, we found that G-MDSCs overexpressed 20 microRNAs and that no microRNAs were lower in expression levels than their counterparts. Among the 20 overexpressed microRNAs in G-MDSCs, miR-486, miR3107 and miR-192 were also overexpressed in M-MDSCs. However, overexpression of miR-486 in G-MDSCs was not validated using RT-PCR. 19 The differential microRNA expression profiles between G-MDSCs and M-MDSCs indicated that G-MDSCs and M-MDSCs share some common microRNAs while also exhibiting distinct microRNAs.
Unfortunately CD11b + Ly6G -Ly6c hi to induced myeloid cell differentiation into MDSCs in vitro, which was reported by Li and his colleagues. 13 In their study, they found that this culture could induce expression of miR-155 and miR-21 as well as the immunosuppression function of MDSCs, which mirrors the profile of MDSCs in vivo induced by tumor. 13 In their study, they did not observe high expression of miR-486 in these cultured cells, which included G-MDSCs and M-MDSCs, using microarrays, in contrast to our results in M-MDSCs from the spleens of LLC-bearing mice. 13 In this study, we found that overexpression of miR-486 in myeloid cells promoted proliferation and suppressed apoptosis and differentiation of these cells.
The target genes of miR-486 that might regulate the proliferation and differentiation of myeloid cells were predicted by bioinformatics analyses. Cebpa was predicted to be a major target gene that could regulate the proliferation of TM-MDSCs. In our study, the expression of Cebpa was inversely correlated with miR-486 in TM-MDSCs measured by RT-qPCR, indicating that TM-MDSCs express a lower level of Cebpa. Hegde, and his colleagues also found that both M-MDSCs (CD11b + Ly6G − Ly6C + ) and G-MDSCs (CD11b + Ly6G + Ly6C + ) induced by 9-Tetrahydrocannabinol (the major bioactive component of marijuana) from peritoneal exudates of mice expressed lower levels of Cebpa than did bone marrow precursors and that Cebpa might be regulated by miR-690. 22 In our study, the results from both luciferase reporter assays in 293T cells and overexpression of miR-486 in myeloid cells indicated that miR-486 inhibits expression of Cebpa, implying that Cebpa is a direct target of miR-486.
A previous study demonstrated that the Cebpa gene is located on chromosome 19q13.1, belonging to the CCAAT/ enhancer-binding protein family, which act as a group of transcription factors with a pivotal role in the differentiation and proliferation of various cell types, including hematopoietic cells. 23 Human Cebpa mRNA exists in myeloid cells but not in lymphoid cells. The expression of the human Cebpa gene starts at the time of stem cell differentiation into the myeloid lineage. 23 Human Cebpa mRNA and protein were found to increase with granulocytic differentiation and to decrease with monocytic differentiation. 24 Overexpression of Cebpa in myeloid cells suppresses proliferation and differentiation and promotees the apoptosis of these cells. The functions of Cebpa in suppressing cell proliferation and promoting apoptosis found in our studies were in accordance with published data. Down-regulating Cebpa by CUGBP1 was reported to promote cell proliferation and suppress apoptosis in human non-small cell lung cancers. 25 Downregulation of Cebpa by IL-17 promotes keratinocyte proliferation. 26 It was also reported that the transcription factor Cebpa is a strong inhibitor of cell proliferation through cyclin-dependent protein kinase 2 (cdk2) and cdk4. 27 Overexpression of Cebpa significantly decreased cell proliferation and upregulated the expression levels of the apoptosis-related genes Foxo3 and Bim in K562 cells. 28 Cebpa siRNA was reported to upregulate iNOS and suppress apoptosis of hepatocytes. 29 Previous research indicated that conditional expression of Cebpa alone in stably transfected bipotential (granulocytic/monocytic) myeloid cell line triggered neutrophilic differentiation, while blocking monocytic differentiation. 17 In our study, overexpression of Cebpa suppressed the differentiation of myeloid cells but had no effect on the proportion of G-MDSCs and M-MDSCs. It was reported that Cebpa and Cebpb play pivotal and partially overlapping roles in determining airway epithelial differentiation. 30 Cebpa suppresses macrophage differentiation by activating QKI-5 and then down-regulating CSF1R expression. 31 Recently, Mackert and his colleagues also discovered that Cebpa expression was significantly reduced in MDSCs from tumor-bearing mice compared to tumor-free animals. Tumor-conditioned medium could down-regulated Cebpa expression. Myeloid lineage specific deletion of Cebpa resulted in significantly enhanced MDSC proliferation and expansion, as well as an increase in myeloid progenitors and a decrease in mature cells. In addition, tumors from Cebpa myeloid conditional null mice have increased MDSC infiltration and angiogenesis in vivo. 32 Findings from Mackert and colleagues were coincident with ours. They also found that down regulation of Cebpa in myeloid cells blocks differentiation, which consistent with our findings that overexpression of miR-486 suppresses the differentiation of myeloid cells by decreasing the expression of Cebpa. However, when we overexpressed Cebpa in myeloid cells, the differentiation of myeloid cells was still suppressed. 31 In addition, Mackert and colleagues found that deletion of Cebpa in MDSCs enhanced the pro-angiogenic, immunosuppressive and pro-tumorigenic behavior of these cells by upregulating the production of iNOS and arginase, as well as MMP-9 and VEGF. 32 Our data show that suppressing the expression of Cebpa by miR-486 and overexpressing of Cebpa inhibits the differentiation of myeloid cells, which seems inconsistent and demonstrates that these regulatory networks are complicated. Furthermore, there may be other target genes whose relationship with miR-486 might be more crucial to cell differentiation, or Cebpa may exert bidirectional regulatory effects on the differentiation of myeloid cells. That is, both overexpression and depletion of Cebpa suppress the differentiation of myeloid cells. The mechanism of miR-486-and Cebpainduced suppression of myeloid cell differentiation needs further study.

| CONCLUSION
Our study indicates that miR-486 promotes proliferation and suppresses apoptosis of myeloid cells by targeting Cebpa, which suppresses differentiation by other unknown target genes that need to be further elucidated. Additionally, miR-486 and Cebpa might be involved in TM-MDSC expansion in tumor-bearing mice.