Osteoblasts secrete miRNA-containing extracellular vesicles that enhance expansion of human umbilical cord blood cells

Osteolineage cells represent one of the critical bone marrow niche components that support maintenance of hematopoietic stem and progenitor cells (HSPCs). Recent studies demonstrate that extracellular vesicles (EVs) regulate stem cell development via horizontal transfer of bioactive cargo, including microRNAs (miRNAs). Using next-generation sequencing we show that human osteoblast-derived EVs contain highly abundant miRNAs specifically enriched in EVs, including critical regulators of hematopoietic proliferation (e.g., miR-29a). EV treatment of human umbilical cord blood-derived CD34+ HSPCs alters the expression of candidate miRNA targets, such as HBP1, BCL2 and PTEN. Furthermore, EVs enhance proliferation of CD34+ cells and their immature subsets in growth factor-driven ex vivo expansion cultures. Importantly, EV-expanded cells retain their differentiation capacity in vitro and successfully engraft in vivo. These discoveries reveal a novel osteoblast-derived EV-mediated mechanism for regulation of HSPC proliferation and warrant consideration of EV-miRNAs for the development of expansion strategies to treat hematological disorders.

for appreciating the complexity of the HSPC-osteolineage-cell crosstalk and may open new avenues for clinical applications.
In the present study, we elucidate the miRNA profile of EVs secreted from human pre-osteoblasts using next-generation sequencing. Based on in silico target prediction analyses and in vitro biochemical analyses we define candidate hematopoietic development pathways affected by osteoblast-EVs. Importantly, we investigate the potency of osteoblast-EVs to promote ex vivo expansion of umbilical cord blood (UCB)-derived CD34 + HSPCs. We further verify the functionality of the expanded cells in vivo by performing xenogeneic transplantation in immunodeficient mice. Our findings provide a foundation for the utilization of EVs as novel tools to modulate hematopoiesis for the development of suitable strategies to treat hematological disorders.

Results
Human osteoblasts secrete EVs that contain small RNAs. To characterize osteoblast-derived EVs, SV-HFO cells were cultured for 12-14 days, and EVs were isolated from the conditioned medium by a series of ultracentrifugation steps. Transmission electron microscopy (Fig. 1a) and nanoparticle tracking analysis (Fig. 1b) show the heterogeneous morphology of the EV population with an average size of 158 nm. Agilent Bioanalyzer RNA profiles show that osteoblast-EVs lack the typical cellular rRNAs, and instead are enriched with small RNAs (Fig. 1c). The EV-RNA peak is retained when the EVs are treated with RNase A prior to RNA isolation (Fig. 1d), verifying that the majority of the detected RNA is indeed present inside the EVs.
To demonstrate that small EV-RNAs comprise miRNAs, we performed quantitative real-time PCR (qPCR) of the widely expressed human miR-1 and miR-24. As shown in Fig. 1e, osteoblast-EVs are devoid of miR-1 (left panel) while they contain relatively high amounts of miR-24 (right panel). Interestingly, RNase A treatment does not significantly alter the miR-24 level, confirming the presence of nuclease-resistant miR-24 inside the EVs. In contrast, exogenously added synthetic miR-1 was immediately degraded when spiked into the sample. Collectively, these data demonstrate that miRNAs are present in the heterogeneous population of osteoblast-EVs. Osteoblast-EVs contain a specific set of abundant miRNAs. To evaluate the miRNA profile of osteoblast-EVs, we performed next-generation sequencing using total RNA from three independent osteoblast cultures and their corresponding EVs. The majority of the miRNAs yielded similar reads counts in cells and EVs (Fig. 2a). In total, we identified 761 mature miRNAs, of which 496 miRNAs are shared between cells and EVs and listed in ExoCarta (Fig. 2b) 26,27 . Notably, all miRNAs exclusively present in cells (82 miRNAs) or EVs (92 miRNAs) are typically of low abundance. Only 185 EV-miRNAs are significantly abundant with levels above Venn diagram shows the number of miRNAs detected in osteoblasts and EVs in comparison with ExoCarta. Numbers in brackets denote the number of highly abundant miRNAs with reads greater than 100 RPKM among all EV miRNAs. (c) Pie chart shows the normalized read count proportions of all EV-miRNAs. Only the top 15 most abundant EV-miRNAs, including let-7f-5p, let-7i-5p, let-7g-5p and let-7a-5p grouped as let7 family, are displayed. (d) Validation of highly abundant EV miRNAs by TaqMan qPCR miRNA assay. Data is presented as raw Ct values (mean ± SD) (N = 3). (e) Volcano plot (significance versus fold change) shows the significantly (P < 0.05, compared with osteoblasts by Student's t-test) abundant (≥ 2-fold; red) and depleted (≤ 0.5-fold; green) miRNAs in EVs. Numbers in brackets denote the number of differentially expressed miRNAs. FC, fold change. (f) IPA network map shows the predicted biological functions significantly targeted by some of the highly abundant (RPKM > 100) enriched EV miRNAs. Blue, EV miRNAs; Green, biological functions.
Scientific RepoRts | 6:32034 | DOI: 10.1038/srep32034 100 normalized reads per million. This set encompasses 183 miRNAs that are among the 496 common miRNAs in both cells and EVs, as well as 2 miRNAs that are not yet listed in ExoCarta. The top 15 of miRNAs present in EVs produce ~75% of all the EV-miRNA reads, and more than half of the reads belong to just three major miRNA families related to miR-21-5p, miR-100-5p and let-7 (let-7f-5p, let-7i-5p, let-7g-5p, let-7a-5p) (Fig. 2c). Figure 2d shows the qPCR validation of selected abundant EV-miRNAs.
Next, we performed relative quantitation of cellular and vesicular miRNAs to assess whether a unique set of miRNAs is enriched in EVs. The Volcano plot in Fig. 2e shows all EV-miRNAs with statistical significance (P < 0.05) and differential abundance (≥ two-fold enrichment); the low abundance miRNAs unique to EVs and cells were excluded from this analysis. The results essentially reveal that EVs are preferentially loaded with 82 miRNAs (red dots) and depleted of 38 miRNAs (green dots). Table 1 shows the highly abundant and enriched 33 miRNAs with normalized reads above 100 ( Supplementary Fig. S1). Among these miRNAs, miR-146a-5p and miR-29a-3p are the most abundant, while miR-1246 and miR-1290 are most enriched (> 10-fold) in EVs. Interestingly, 21 out of 33 enriched EV-miRNAs have not been previously reported as EV markers that are most commonly detected across all tissues and body fluids 28 . Hence, these 21 miRNAs may reflect selective sorting into osteoblast-EVs.
Ingenuity Pathway Analysis (IPA) was performed to predict the impact of enriched EV-miRNAs on target cell gene expression and phenotype. The most significantly annotated molecular functions are cellular development, cellular growth and proliferation, cell cycle, cell death and survival and cellular compromise (Fig. 2f). Together, our comparative analyses revealed overrepresentation of a selective group of miRNAs in osteoblast-EVs, indicating specific packaging of biologically functional regulatory molecules as EV cargo.
Osteoblast-EVs are enriched with miRNAs crucial for hematopoiesis. Each stage of hematopoietic differentiation is characterized by a specific miRNA signature 29 . We analyzed the osteoblast-EV-miRNA content for known regulators of hematopoiesis to find cues for EV-mediated HSPC-osteolineage-cell crosstalk. Indeed, a number of miRNAs that may control distinct stages of hematopoietic development are detected in high abundance in EVs (Fig. 3a). Hence, we investigated whether treatment of HSPCs with osteoblast-EVs would alter expression of specific mRNAs that are targeted by miRNAs enriched in osteoblast-EVs. We focused initially on miR-29a, which is significantly enriched in osteoblast-EVs, because it is known to be involved in early stages of hematopoiesis as a regulator of self-renewal, survival and proliferation of HSCs and HSPCs 30 . In silico target prediction analysis by TargetScan combined with supervised literature searches refined the large list of potential targets to a defined list of validated miR-29a target genes relevant to HSPCs. By qPCR analysis we evaluated the expression level of miR-29a target genes involved in proliferation (TET2, PTEN), apoptosis (BCL2), cell cycle regulation (HBP1, CDC42EP2) and extracellular matrix adhesion (COL1A1, ELN) of HSPCs [30][31][32] . Osteoblast-EVs reduce the expression of all selected target genes, except TET2, in human UCB-derived CD34 + HSPCs (Fig. 3b). EVs significantly down-regulate the expression of HBP1, BCL2 and PTEN, suggesting that osteoblast-EVs control proliferation of the recipient HSPCs. Proliferation assays that monitor Ki-67 staining (Fig. 3c) and cell cycle analysis based on DNA content (Fig. 3d) establish that osteoblast-EVs stimulate proliferation and cell cycle progression of the CD34 + cells 24 hours after treatment. Thus, the integrated results of both in silico and in vitro approaches indicate that osteoblast-EVs are enriched with miRNAs involved in signaling cascades that regulate HSPC proliferation.
Osteoblast-EVs promote ex vivo expansion of CD34 + HSPCs. We evaluated the capacity of osteoblast-EVs to promote the ex vivo expansion of human UCB-derived CD34 + cells in growth factor (stem Expression levels (mean ± SD) of candidate mRNA targets of miR-29a in human UCB-derived CD34 + cells after 24 hours of incubation in GBGM supplemented with SCF and Flt3L in the absence (control) or presence of osteoblast-EVs (N = 5). *P < 0.05, ***P < 0.005, compared with control by Student's t-test. (c) Proliferation assay, determined by the % of Ki-67 + CD34 + cells (mean ± SD), after 24 hours incubation in the absence (control) or presence of osteoblast-EVs (N = 3). *P < 0.05, compared with control by Student's t-test. (d) Cell cycle distribution, using propidium iodide (PI) staining of CD34 + cells (mean ± SD), after 24 hours incubation in the absence (control) or presence of osteoblast-EVs (N = 3). **P < 0.01, compared with control by two-way ANOVA.
Scientific RepoRts | 6:32034 | DOI: 10.1038/srep32034 cell factor, SCF and Fms-related tyrosine kinase 3 ligand, Flt3L)-driven serum-free expansion cultures. Osteoblast-EVs induce a two-fold expansion of both total number of viable nucleated cells (TNCs) (P < 0.01) and CD34 + cells (P < 0.01) in 10 days as compared to control cultures (Fig. 4a,b). We also assessed the presence of the most immature CD34 + cell subset in the expanded cells by multicolor flow cytometry using markers for primitive HSPCs (lin -CD34 + CD38 low CD45RA low CD90 + ), referred to as phenotypic HSCs. EV treatment also increases the number of phenotypic HSCs more than two-fold (P < 0.005) compared to the control on day 10 (Fig. 4c). Interestingly, the expansion potential of EVs remains clearly evident even in the presence of other highly effective expansion factors, such as StemRegenin 1 (SR1) and thrombopoietin (TPO) 33,34 (Supplementary Fig. S2).
Next, we investigated whether the CD34 + cells that were expanded ex vivo using EVs retain their differentiation capacity in vitro by performing colony-forming unit (CFU) assay. EV-expanded cells exhibit a higher clonogenicity, most likely due to the increased number of viable and functional CD34 + cells after expansion (Fig. 4d). However, the frequencies of multilineage progenitors (CFU-GEMM), erythroid progenitors (CFU-E/BFU-E) and granulocyte/macrophage progenitors (CFU-G/M/GM) remain comparable to the control, indicating that EVs promote expansion but do not favor specific hematopoietic lineages. These findings demonstrate the potency of osteoblast-EVs to promote growth factor-driven HSPC expansion while retaining the pool of progenitor cells that give rise to erythrocytes and myeloid cells in vitro.

Osteoblast-EV-expanded cells retain in vivo engraftment potential.
To assess the impact on engraftment and hematopoietic repopulating ability of the ex vivo expanded cells, sublethally irradiated immunodeficient NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJ (NSG) mice were transplanted with CD34 + progeny cells derived from 10 5 seeded cells after expansion with or without EVs for 10 days. Engraftment was defined as at least 0.1% human chimerism in the peripheral blood. All mice show similar levels of human chimerism (~20%) in both control and EV treatment groups at 19 weeks after transplantation (Fig. 5a). EV treatment does not alter the speed and quality of recovery of different human lymphoid and myeloid lineages in NSG mice. The predominant human cell population after engraftment consists of CD19 + B-cells, as well as very low levels of other lymphoid (NK and T-cells) and myeloid cells (Fig. 5b). Moreover, both control and the EV treatment groups exhibit similar levels of CD45 + human chimerism in recipient bone marrow at 21 weeks after transplantation (Fig. 5c). Similar to the peripheral blood, the majority of human cells in the bone marrow expressed the B-cell marker CD19 (Supplementary Fig. S3). As expected, only a small (2%) fraction of human cells is positive for CD34 (Fig. 5d). We note that the mice in the control group show a high degree of variability (C V : 54.13%, Fig. 5c; C V : 100.75%, Fig. 5d) of chimerism in the bone marrow, possibly explained by the lower number of total transplanted cells after expansion. Taken together, these findings clearly demonstrate that ex vivo EV treatment retains the engraftment potential of human cells in NSG mice.
Osteoblast-EVs stimulate the proliferation of immature cells. Most conventional expansion protocols, which provide short-term robust proliferation of the CD34 + progenitor cells, are accompanied by concomitant differentiation and result in loss of primitive HSC sub-populations [35][36][37] . To study the effect of EVs on immature stem cells, we sorted phenotypic HSCs as a starting population. Osteoblast-EV treatment of phenotypic HSCs significantly induces the expansion of TNCs (Fig. 6a), CD34 + cells (Fig. 6b) and phenotypic HSCs (Fig. 6c) after 10 days. Corroborating these results, CFU assays reveal an increase in the number of the immature cells while retaining the frequency of the different lineages (Fig. 6d).
To determine the effect of osteoblast-EVs on the maintenance of CD34 and CD90 expression after successive cell divisions of phenotypic HSCs, we used CellTrace TM Violet staining. Cells that were grown in the presence of osteoblast-EVs undergo a higher number of cell divisions while keeping their stem cell phenotype after 5 days compared to control (Fig. 7a,b). Figure 7c,d show the fraction of cells that have divided up to 4 times within the CD34 + cell population in the absence or presence of EVs. Our data show that EVs stimulate the cell division kinetics of the immature cells, resulting in an increased number of CD34 + cells.

Discussion
EV-mediated intercellular communication is an exciting new area of research that is rapidly evolving thanks to the emergence of powerful tools that enable characterization of their bioactive cargo. We used next-generation sequencing to study the miRNA profile of human osteoblast-derived EVs. Based on the overrepresented EV-miRNAs we delineated the targeted biological functions associated with hematopoiesis and verified in silico predictions with in vitro results. We show that osteoblast-EVs stimulate the proliferation of UCB-derived CD34 + HSPCs and subsequent expansion of functional primitive cells, which retain their multipotency in vitro and successfully engraft in vivo. Our findings provide a fundamental description of the biological roles of EVs in mediating crosstalk between osteolineage cells and HSPCs. Equally important, osteoblast-derived EVs may present potential value in clinical applications to treat hematopoietic disorders.
There is considerable interest in understanding how circulating miRNAs exert their effect in a paracrine manner via EV-mediated delivery to the target tissue. We performed high throughput analysis of miRNAs to assess whether human osteoblast-EVs utilize small RNAs to communicate with HSPCs in the bone marrow niche. As expected, the EV-miRNA profile mirrors that of the osteoblastic donor cell, but we also detected a large number of miRNAs that appear to be enriched in EVs. Importantly, we identified a number of miRNAs that are abundant in osteoblast-EVs and are known regulators of hematopoiesis. To predict targeted pathways, we focused on miR-29a because it controls early steps of hematopoiesis and is enriched in EVs compared to the donor cells. Our study shows that osteoblast-EVs are capable of down-regulating the expression of cell cycle-and growth-related miR-29a target genes in CD34 + cells. We further demonstrate that osteoblast-EVs stimulate proliferation and cell cycle progression of CD34 + cells, in accordance with the expression profile of down-regulated target genes. Further investigation is required to elucidate the precise mechanism of EV action by showing a direct evidence of delivery of miR-29a via EVs, followed by miR-29a regulation of the targets. Nevertheless, attributing the proliferative effect of EVs solemnly to one miRNA would be oversimplification of the complexity of EV-mediated cell-to-cell communication. In fact, our previous studies reporting the discovery of osteoblast-EV proteins capable of stimulating cell growth demonstrates the necessity of carefully investigating other functional cargo to understand the EV effect.
Stem cells are special cells with remarkable regenerative abilities, which make them very attractive for the development of cell-based therapies. A complex network of environmental cues tightly regulates stem cell fate determination, including self-renewal and differentiation. Ratajczak and colleagues were one of the firsts to report the role of EVs in stem cell regulation 38 . They showed that embryonic stem cells released EVs containing regulatory proteins and mRNA capable of reprogramming HSCs. Since then numerous studies reported the role of EVs in stem cell biology with a major focus on miRNA transfer, suggesting a critical paracrine role for EVs in stem cell niches [39][40][41][42] . We previously showed that osteoblast-EVs induced proliferation of bone-metastasizing prostate cancer cells 22 . Recent studies indicate that prostate cancer cells compete with HSCs for the bone marrow niche, emphasizing the importance of EVs in the regulation of HSC growth and survival 43 . Hematopoietic differentiation is characterized by specific miRNA signatures, which act in each step of linage decision to ensure proper hematopoiesis 25,29 . The osteoblast-EV-miRNA profile we describe here provides further insights into the complexities of the HSPC-osteolineage-cell crosstalk.
The ability of osteoblast-derived EVs to regulate HSPCs brings exciting possibilities for therapeutic applications. UCB is an increasingly important source of HSPCs for the treatment of a variety of hematological disorders. However, UCB grafts contain a low number of HSPCs that poses as a limiting factor for proper engraftment, which leads to delayed hematopoietic recovery and patient morbidity and mortality 44,45 . Over the past few decades, studies have been focusing on the establishment of ex vivo culture systems to expand UCB-derived HSPCs for improved engraftment and post-transplantation recovery 33,36 . Therefore, we addressed whether we could therapeutically exploit the proliferation promoting capacity of osteoblast-EVs to expand UCB-derived CD34 + cells ex vivo. Conventional expansion protocols make use of a cocktail of growth factors to provide culture conditions optimized for HSPC expansion to obtain short-term robust proliferation 33,46 . Here, we limited the combination of growth factors to the minimum required potency to better study the EV effect. Our findings show that EVs significantly enhance the ex vivo expansion of CD34 + cells, including the most primitive subset of phenotypic HSC. Most importantly, EV-expanded CD34 + cells differentiate in vitro and successfully engraft and re-populate NSG mice in vivo. Further integrated genomic and proteomic analyses are of utmost importance to unravel candidate EV cargo components that can be modulated to increase expansion capacity that will meet the clinical need.
In conclusion, our findings provide a new paradigm for the role of EVs in the regulation of stem cell niches. We propose that EVs contain a signature miRNA profile that participates in the HSPC-osteolineage-cell crosstalk. It remains to be determined whether EV-miRNAs act in concert with other regulatory EV cargo, such as mRNAs and proteins, which may contribute to the proliferative effect of osteoblast-EVs. Identification of such critical EV components opens up avenues to be exploited clinically to develop novel approaches not only for the treatment of hematological disorders but in a broader context also for use in regenerative medicine.
EV isolation and characterization. Osteoblast-EVs were isolated from 20 ml conditioned medium by low speed centrifugation (1500 rpm, 5 minutes; 4500 rpm, 10 minutes) followed by ultracentrifugation (20,000 g, 30 minutes; 100,000 g, 1 hour at 4 °C) of the supernatant using the SW32Ti rotor (Beckman Coulter, Fullerton, CA, USA). EVs were prepared as 100 μ l suspensions, and the amount of experimental EV dose was determined as 5% (v/v). Transmission electron microscopy images were taken as previously described 22 . EV size distribution and concentration was measured with NanoSight LM10 (Nanosight Ltd., Amesbury, UK) equipped with a 405 nm laser. Each sample was tracked for 60 seconds with 3 repetitions. The data was processed by NTA 2.3 software.
RNA isolation and quantitative real-time PCR. The purified EV pellet was incubated with or without RNase A (100 mM) and synthetic miR-1 (20 pM) at 37 °C for 30 minutes, and total RNA was isolated using the TRIzol reagent (Thermo Fisher Scientific) according to the manufacturer's instructions. RNA concentration was determined using Nanodrop (Thermo Fisher Scientific) and size distribution was checked on an Agilent Bioanalyzer RNA 6000 Pico chip (Thermo Fisher Scientific). RNA from CD34 + cells was isolated using NucleoSpin RNA XS kit (Macherey-Nagel, Duren, Germany) according to the manufacturer's instructions.
Quantitative real-time PCR for mRNAs and miRNAs were performed using the SYBR ® Green (Eurogentec, Seraing, Belgium) and Taqman ® kits (Thermo Fisher Scientific), respectively, according to the manufacturer's instructions. The primer sequences are listed in Supplementary Table S1.
Next-generation sequencing and bioinformatic analysis of miRNAs. Sequencing of miRNAs was performed by Illumina HiSeq with samples prepared with the NEBNext Small RNA library preparation kit. We used the CAP-miRSeq bioinformatic pipeline for mapping of miRNA reads data analysis 48 . In brief, adaptor sequences were removed from 50 bp reads using Cutadapt 49 and sequences of sufficient size (> 17 nucleotides in length) were aligned to the hg19 reference genome and miRBase 19 reference sequences using Bowtie 50 . Quantitative analysis of known and predicted miRNAs was performed using miRDeep2 51 . Sequencing data are Scientific RepoRts | 6:32034 | DOI: 10.1038/srep32034 available in the ArrayExpress database (www.ebi.ac.uk/arrayexpress) under accession number E-MTAB-4560. Comparative analyses of the EV data were obtained by querying the ExoCarta (V5, 29 July 2015) database. IPA (Ingenuity ® Systems, www.ingenuity.com) and TargetScan (Release 7.0) were used to predict target pathways and genes.
Flow cytometry. Analysis of proliferation and cell cycle progression of CD34 + cells was performed using flow cytometry by monitoring the proliferation marker Ki-67 relative to DNA content. Cells were incubated with or without EVs for 24 hours, and stained with Alexa Fluor ® 488-conjugated Ki-67 antibody (BD Biosciences, San Jose, CA, USA) followed by the addition of propidium iodide (BD Biosciences) according to the manufacturers instructions. All samples were immediately analyzed using BD Accuri TM C6 flow cytometry (BD Biosciences), and the data was analyzed using C6 software. Flow cytometry was also used to examine subpopulations of CD34 + cells. Absolute numbers of viable human CD34 + cells were determined by a single platform flow cytometric assay using anti-FITC-CD45, anti-CD34-PE, fluorescence reference counting beads (Beckman Coulter) and DAPI (Sigma). The frequencies of human phenotypic HSCs were determined using anti-Lin-FITC, anti-CD38-APC, anti-CD90-PE (eBiosciences, Vienna, Austria), anti-CD34-PE-Cy7, anti-CD45RA-APC-H7 (BD Biosciences); and DAPI. Human chimerism in mice blood and bone marrow was determined using anti-mouse CD45-eFluor450, (eBiosciences) and anti-human CD45-APC-Cy7 (BioLegend, London, UK). The human lymphoid and myeloid lineages were analyzed by staining with anti-CD19-PE, anti-CD3-APC, anti-CD33-PerCP-eFluor710, anti-CD56-PE-Cy7, and anti-CD15-BV510. All samples were analyzed using BD FACSCantoII (BD Biosciences), and the data was analyzed using FlowJo software (Tree Star, Inc., Ashland, OR, USA). CFU assays. Suspensions of 300 expanded CD34 + cells in MethoCult ™ medium (MethoCult ™ GF H4434, Stemcell Technologies, Vancouver, BC, Canada), were incubated in standard culture dishes for 11 days at 37 °C in a humidified atmosphere of 5% CO 2 . The colonies were counted using an inverted light microscope, and the different lineages were evaluated by morphology according to the manufacturer's instructions (Stemcell Technologies).
Transplantation. NSG mice were sub-lethally irradiated (3 Gy), and intravenously transplanted with the progeny generated from 10 5 CD34 + cells cultured with or without EVs for 10 days. Peripheral blood was collected at 2 weeks intervals starting at 3 weeks post-transplantation through a small incision in the tail vein. At 21 weeks post-transplantation, the mice were sacrificed by cervical dislocation, and the bone marrow was harvested from the femurs as previously described 52 .