EMBRYONIC STEM CELLS/INDUCED PLURIPOTENT STEM CELLS

Transcription

1 EMBRYONIC STEM CELLS/INDUCED PLURIPOTENT STEM CELLS MicroRNA-200C and -150 Play an Important Role in Endothelial Cell Differentiation and Vasculogenesis by Targeting Transcription Repressor ZEB1 ZHENLING LUO, a,b GUANMEI WEN, a,c GANG WANG, d XIANGYUAN PU, a SHU YE, a QINGBO XU, d WEN WANG, b * QINGZHONG XIAO a * a Centre for Clinical Pharmacology, William Harvey Research Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, United Kingdom; b Institute of Bioengineering, Queen Mary University of London, London, United Kingdom; c Department of Pathophysiology, Guangzhou Medical University, Guangzhou, People s Republic of China; d Cardiovascular Division, British Heart Foundation Centre, King s College London, London, United Kingdom Key words: MicroRNAs mir-200c mir-150 Stem cells Endothelial cells Cell differentiation Zinc finger E-box-binding homeobox 1 Vasculogenesis ABSTRACT To investigate the role of mirna in controlling human embryonic stem (hes) cell differentiation toward the endothelial lineage and chick embryonic blood vessel formation, undifferentiated hes cells were first cultured on Matrigelcoated flasks and in endothelial cell growth medium-2 (EGM-2) to initiate endothelial cell (EC) differentiation. CD146 1 cells were isolated from differentiating hes cells and expanded in vitro. The in vitro expanded CD146 1 cells were positive for EC markers, capable of Ac-LDL uptake, lectin binding, and the formation of vascular structures in vitro and in vivo. mirna gain/loss-of-function analyses revealed that mir-150 and mir-200c were crucial in EC differentiation. Transcriptional repressor zinc finger E-boxbinding homeobox 1 (ZEB1) was identified as the communal Disclosure of potential conflicts of interest is found at the end of this article. target gene of mirna-200c and 2150, and inhibition of ZEB1 was required for mirna-200c or 2150 mediated EC gene expressions. Moreover, we demonstrated that ZEB1 could transcriptionally repress EC gene expression through direct binding to promoters of EC genes. Finally, we also demonstrated that mirna-200c and 2150 played an important role in chick embryonic blood vessel formation by in vivo inhibition of mirna-200c or 2150 in developing chick embryos, and blocking ZEB1 signaling in CD146-positive cells could rescue the inhibitory effects of mir-200c inhibiton in in vivo vasculogenesis. Our findings revealed that mir-150 and mir-200c play an important role in human endothelial lineage specification and chick embryonic vasculogenesis by targeting ZEB1. STEM CELLS 2013;31: INTRODUCTION The use of stem cells in vascular regeneration has gained immense momentum over the years. Various types of stem cells, particularly embryonic stem (ES) cells, have been under investigation for use in regenerative medicine because these stem cells hold significant potential for clinical therapies due to their distinctive capacity to both self-renew and differentiate into a wide range of specialized cell types including vascular endothelial cells (ECs) [1,2] and smooth muscle cells (SMCs) [3 9]. EC differentiation is an initial step of vasculogenesis which plays critical roles during embryonic and postnatal cardiovascular development. Moreover, ECs derived from stem/progenitor cells provide an unlimited cell source for cell therapy in various cardiovascular diseases, such as vascular injury, stroke, coronary heart disease, and acute myocardial infarction. Understanding the transcriptional regulatory circuitry of EC differentiation is fundamental in understanding human cardiovascular system development and realizing the therapeutic potential of these cells. MicroRNAs (mirnas) are endogenous, highly conserved, short noncoding 22 nucleotide RNAs, and constitute a novel class of gene expression regulators which play important roles in various aspects of development, homeostasis, and disease [10,11]. Traditionally, mature mirnas are thought to suppress gene expression by inducing mrna cleavage [12,13] or mrna decay [14,15] or by inhibiting mrna translation [16]. Author contributions: Z.L., G.W., G.W., and X.P.: collection and/or assembly of data; S.Y.: manuscript writing and/or provision of study materials; Q.X. and W.W.: conception and design and financial support; Q.X.: conception and design, financial support, administrative support, provision of study material, collection and/or assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript. Z.L. and G.W. contributed equally to this article. Correspondence: Qingzhong Xiao, M.D., Ph.D., Centre for Clinical Pharmacology, William Harvey Research Institute, Barts and the London School of Medicine and Dentistry, John Vane Science Centre, Charterhouse Square, London EC1M 6BQ, United Kingdom. Telephone: ; Fax: ; q.xiao@qmul.ac.uk; or Wen Wang, Ph.D., Institute of Bioengineering, Queen Mary, University of London, Mile End Road, London E1 4NS, United Kingdom. Telephone: ; Fax: ; wen.wang@qmul.ac.uk Received December 27, 2012; accepted for publication May 6, 2013; first published online in STEM CELLS EXPRESS June 14, VC AlphaMed Press /2013/$30.00/0 doi: /stem.1448 STEM CELLS 2013;31:

2 1750 MicroRNAs in Endothelial Cell Differentiation However, recent evidence suggests that mirnas can also target the 5 0 -untranslated regions (UTRs) or coding regions [17] of their target genes and, in some cases, can upregulate gene expression [11]. Interestingly, many mirnas are expressed in a tissue-specific manner, suggesting that certain mirnas might be important for cell/tissue specification [18]. An approach to examine the spectrum of the biological significance of mirnas in stem cell differentiation is to remove all mirnas in undifferentiated ES cells by mutation or disruption of Dicer and/or Drosha, the rate-limiting enzyme involved in the maturation of mirnas. By using this strategy, previous studies have suggested that mirnas play important roles in ES cell self-renewal and differentiation [19,20]. An essential role of mirnas in cardiovascular development has been demonstrated in a study of Dicer-deficient mice which showed that the loss of mirnas resulted in severe impairment of heart and blood vessel development [21]. Studies aimed at elucidating the role of individual mirnas in the regulation of vascular system formation and vascular diseases were increasingly being performed globally, but many were focused on the functional effects of mirnas in regulating EC angiogenic phenotypes such as migration, proliferation, and/or morphogenesis [22 26]. So far, only a few most recent studies suggest a role of individual mirna in EC differentiation from ES cells or EC gene regulation, such as has-mir-99b, 181a and 181b [27], has-mir-6086/6087 [28], and has-mir [29]. However, the significance and exact role of individual mirnas in EC differentiation and cardiovascular system development remains to be further elucidated. In this study, we have demonstrated for the first time that mir-150 and mir-200c play an important role in hes cell differentiation toward the EC lineage and chick embryonic vasculogenesis by targeting ZEB1, a potential transcriptional repressor of EC genes. MATERIALS AND METHODS A detailed description on the materials and methods is provided in the supporting information. Briefly, the Shef hes cell lines were obtained from the United Kingdom Stem Cell Bank (UKSCB, Hertfordshire, U.K. and routinely maintained in our Laboratory according to protocols provided by UKSCB. Undifferentiated hes cells were cultured on Matrigel-coated flasks and in EGM-2 to initiate EC differentiation. Fluorescence-activated cell sorting was used to isolate CD146-positive cells from differentiating hes cells. Sorted CD146-positive cells were resuspended in human ES cell medium for longer term of maintenance or EGM-2 and cultured onto collagen I-coated dishes for EC differentiation and in vitro expansion for 5 10 passages (15 35 days). Real time quantitative PCR (RT-qPCR), immunoblotting, flow cytometry, and immunofluorescent staining analyses were carried out to examine the EC gene and protein expression levels, respectively. AcLDL uptake and Lectin binding, in vitro tube formation, and in vivo angiogenesis assays were conducted to characterize the functional characteristics of in vitro expanded ECs derived from hes cells. MicroRNA transfection experiments by using mirna precursors and inhibitors were carried out in the differentiating hes cells to reveal its functional role of respective mirna in EC specification. Human ZEB1/TCF UTR was cloned into pmir-reporterbasic vector for evaluating the regulatory role of mir-200c and 2150 in ZEB1 gene regulation, and mir-200c and mir-150 binding sites mutations within ZEB1/TCF UTR were generated by using QuikChange site-directed mutagenesis kit to further confirm its requirement of these binding sites in ZEB1 gene regulation by these two mirnas. small interfering RNA (sirna) experiments were carried out in the differentiating hes cells to explore its functional role of ZEB1 in EC specification. Different EC gene promoters were cloned into pgl3-enhancer vector and luciferase activity assay was performed to measure respective EC gene promoter activity. Chromatin immunoprecipitation (ChIP) assay was used to detect the direct binding of ZEB1 protein to the promoter region of EC genes. In vivo transfection of mirna inhibitors into developing chick embryos and chorioallantoic membrane assay were performed to evaluate its in vivo significance of mir-200c and 2150 in chick embryonic vasculogenesis. Matrigel-cell implantation with various treatments was conducted to examine the combinatory effects of mir-200c inhibition and ZEB knockdown in in vivo vasculogenesis. RESULTS hes Cell Differentiation Toward the EC Lineage To drive EC differentiation from hes cells, undifferentiated hes cells were plated onto Matrigel Basement Membrane Matrix (BD Biosciences, Bedford, MA, coated dishes or plates and cultured in EGM-2 medium for the indicated experimental time points. Total RNA and protein were extracted and subjected to RT-qPCR and Western blot analysis, respectively. The expression levels of a panel of EC specific gene markers such as CD144/VEcad, vwf, VEGFR2/KDR, CD31/PECAM-1, and enos were significantly upregulated when the period of differentiation increased, as demonstrated by RT-qPCR (Fig. 1A) and/or Western blot (Fig. 1B) analyses. The inductions of protein expression of these EC-specific markers were further confirmed by immunofluorescent staining in day 9 differentiated ECs (Fig. 1C). Moreover, hes cell specific marker TRA-1 60 was strongly expressed in undifferentiated hes cells as expected, while hes cells were negative for EC specific marker CD144 (Fig. 1D). In contrast, compared to day 0 of undifferentiated hes cells, expression levels of stem cell specific transcriptional genes, Nanog and Oct4, were significantly decreased at days 3, 5, 7, and 9 during EC differentiation, and displayed a very low or undetectable level in the fully differentiated ECs (CD146-positive cells, passage 5 10) and human umbilical vein endothelial cells (HUVECs) (Fig. 1E, 1F), suggesting that hes cells were gradually committing toward the EC lineage. CD146 Serves As a Good Selection Marker for EC Generation from hes Cells Previous studies suggest that CD146 is a perivascular and EC marker and was used to identify bone marrow [30] and endometrium [31] mesenchymal stem cells. Our flow cytometry analysis showed that approximately 11.2% of the cell population expressed CD146 at day 7 of differentiation, which reached a peak at 34.6% at day 12 of differentiation (supporting information Fig. S1A). Interestingly, we observed that only a few CD146-positive cells (0.5%) expressed the mature EC marker CD144 at day 7 of differentiation, but many more CD146- positive cells (15.2%) coexpressed CD144 at a later stage of differentiation (day 12) (supporting information Fig. S1B), suggesting that CD146 can be used as a selection marker for endothelial committing cells and/or EC generation from differentiating hes cells. To obtain such endothelial committing cells and further generate large number of more homogeneous ECs from the differentiating heterogeneous hes cells, CD146- positive cells were sorted from day 7 differentiating cells with high purity (96.5%) (supporting information Fig. S1B) and cultured onto collagen I-coated dishes in EGM-2 to allow EC

3 Luo, Wen, Wang et al Figure 1. Differentiation of human embryonic stem (hes) cells toward the endothelial cell lineage. Clumps of undifferentiated hes cells were seeded on Matrigel and cultured in endothelial cell growth medium-2 (EGM-2) for 3, 5, 7, and 9 days. During each respective time point, total RNA and protein were extracted and subjected to real-time polymerase chain reaction analysis (A, E) and Western blot analysis (B, F). Day 0 samples were undifferentiated hes cells and served as negative control. (C): Expression of EC markers such as CD31/PECAM-1, CD144/VEcad, CD146, and vwf in hes cells-derived endothelial cells (ECs). Day 9 differentiated hes cells grown in EGM-2 were probed with primary antibodies (CD31, CD144, CD146, or vwf), followed by polyclonal secondary antibodies conjugated with fluorescein isothiocyanate or tetramethylrhodamine isothiocyanate fluorescence. (D): Human ES cell-specific marker TRA-1 60, but not EC specific marker CD144, expressed in undifferentiated hes cells. Representative fluorescent images from three independent experiments were presented in (C) and (D). (E, F): Expression levels of stem cell specific transcription factors were significantly downregulated during EC differentiation and almost undetectable in the fully differentiated ECs as well as adult HuVECs. The data presented here are representative or mean6s.e.m. of three independent experiments. *, p <.05. Abbreviations: hues, human embryonic stem; HuVECs, human umbilical vein endothelial cells. differentiation and/or maturation. In some cases, sorted CD146- positive cells were cultured in hes cell culturing medium for further in vitro expansion. Early microscopic examinations of the sorted CD1461 cells grown in EGM-2 conditions revealed that two different cell phenotypes were observed (supporting information Fig. S1Ci). Although both cobblestone and spindleshaped cells were present in the early passage of CD1461 cell culture, the spindle-shaped cells were absent and only a population of cells with cobbled-stone morphology was observed as the passage number of the cell culture increased (supporting

4 1752 MicroRNAs in Endothelial Cell Differentiation information Fig. S1Cii). Immunocytochemistry analysis showed that these cells were positive for EC specific markers CD144 and enos with high percentage (75% 94%) at passage 5 10 (supporting information Fig. S1D). Flow cytometry analysis showed that only small fractions (24.5% 6 3.9%) of the freshly isolated CD146-positive cells or sorted CD146-positive cells cultured in stem cell medium (passage 0 3) expressed mature EC marker CD144, while majority of cells (91.3% %) were positive for CD144 when they were cultured in EGM-2 for more than five passages, which are consistent with the immunostaining data (supporting information Fig. S1D), implying that freshly isolated or early passage of CD146-positive cells can be referred as EC committing cells, but the late passage of CD146- positive cells (passage 5 10) cultured in EGM-2 medium can be used as fully differentiated ECs or EC-committed cells. Importantly, by counting and calculating the final EC number from two independent experiments, we found approximately of ECs could be generated from 10 6 of hes cells over 45 days. These data strongly suggest that a large number of ECs can be generated from hes cells. In Vitro and In Vivo Functional Characterization of ECs Derived from hes Cells To further characterize these ECs, in vitro function tests such as Dil-Ac-LDL uptake, lectin binding, and tube formation assays were carried out in this study. As expected, these expanded CD1461 cell population (passage 5 10) were capable of Dil-Ac-LDL uptake and lectin binding (supporting information Fig. S2A). Importantly, when these ECs were seeded onto a Matrigel-coated surface, an early formation of a network and some fine vascular networks were observed at 2 hours and 24 hours since plating, respectively (supporting information Fig. S2B). To further investigate whether these ECs have angiogenic functions in vivo, PKH26-labeled ECs were mixed with Matrigel, then injected into severe combined immunodeficiency mice, and harvested at 2 weeks postprocedure. The Matrigel plugs were snap-frozen in liquid nitrogen, sectioned, and subjected to H&E and immunofluorescence staining. Data shown in supporting information Figure S2C revealed a complex vascular network which had formed in the presence of the transplanted ECs. Importantly, we observed a large number of ECs within Matrigel implants were labeled with red fluorescent (PKH26) (supporting information Fig. S2D), indicating its exogenous origins. The combination of all these assays demonstrated that the in vitro differentiated and expanded CD1461 cell population could function as mature ECs in vitro and in vivo. mirnas Expression During EC Differentiation from hes Cells To identify potential mirna candidates for EC differentiation, total RNA including small RNA were harvested from undifferentiated hes cells (day0) or differentiating hes cells at day 3 (d3), 6 (d6), and 9 (d9) and subjected to microarrays analysis (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany, Data from the microarrays analysis revealed that angiogenic promoting mirnas such as mir-126 [22,23] and mir-210 [32] were initially undetected at days 3 and 6 but surfaced at day 9 of differentiation (supporting information Table S2). In addition, another positive regulator of angiogenesis, mir-130a [24] was found in an increasing trend over the 9-day differentiation period. Moreover, levels of antiangiogenic mir-221 [33,34] appeared to decrease over the differentiation time frame (supporting information Table S2). mir-145, which has been previously reported to facilitate ES cell differentiation by repressing the core pluripotency factors OCT4, sex determining region Y- box 2 (SOX2), and Kruppel-like factor 4 (KLF4) [35], was induced at day 9 as shown in the mirna array analysis. Importantly, almost all the mirnas belong to the previously reported two human ES cell specific mirna clusters [36], such as mirna cluster (mir-302a, mir-302a*, mir-302b, mir-302c, mir-302c*, mir-302d, and mir-367) and mirna cluster (mir-371, mir-372, and mir- 373), were significantly decreased or undetectable during EC differentiation (supporting information Table S2), which were further confirmed by RT-qPCR analyses (supporting information Fig. S3). These observations clearly demonstrate that our differentiation process of hes cells toward the EC lineage does coincide with some of the previously published mirna findings. Interestingly, the microarray analysis showed five mirnas (mir-150, mir-1915 and mir-200 family members, mir-141, mir-200c, and mir-205) were upregulated to a greater magnitude than the rest of detectable mirnas during EC differentiation, which was further confirmed by real-time RT-PCR analysis (Fig. 2A). These findings indicate that these mirnas may play a role in EC differentiation from hes cells. MiR-150 and 200c Are Involved in EC Differentiation from hes Cells To investigate whether these mirnas are involved in EC differentiation, mirna loss-of-function experiments were conducted by using anti-mir hsa-mirnas (141, 150, 200c, 205, and 1915) inhibitors (Ambion) in differentiating hes cells. The experiments showed that mir-150 inhibition reduced the expression of EC-specific genes (vwf, VEGFR2, CD144, and CD146) in the differentiating hes cells (Fig. 2B). Similar data were observed when mir-200c expression in the differentiating hes cells was inhibited by using mir-200c inhibitor (Fig. 2C). Western blot analysis revealed that there was also a decreased protein production of these genes (Fig. 2D). In contrast, no such effects were observed when expressions of mir-1915, 141, and 205 were inhibited in the differentiating hes cells (supporting information Fig. S4). To further explore the functional role of mir-150/200c in EC differentiation, mirna gain-of-function experiments were conducted by using pre-mir hsa-mir-150 or 2200c precursors (Ambion) in differentiating hes cells. Our data showed that the overexpression of mir-150 or 200c significantly upregulated EC specific gene and protein levels (Fig. 3A 3C). Interestingly, no significant changes were observed among SMC-specific gene expression such as smooth muscle a-actin and myosin heavy chain in response to the gene modulation of mir-150 or 200c (data not shown). These suggest that mir-150 and 200c are rather specifically involved in EC differentiation from hes cells. The increase of mrna and protein level induced by mir- 200c or 150 over-expressions in the differentiating cells could result from an increase of the expression in a single ECcommitted cell or the increase of the number of ECcommitted cells or both. To this aim, we first examined if mir-200c or 150 over-expressions will increase ECcommitting cell number. Day 6 differentiating hes cells were transfected with mir precursor negative control, mir-200c or mir-150 precursor, respectively, and cultured in EGM-2 medium for 3 days. We observed that mir-200 c or 150 overexpressions significantly increased CD146-positive cell (EC committing cell as mentioned above) number in the differentiating hes cells (Fig. 3D). To further distinguish the above possibilities, sorted CD146 cells (freshly isolated or passage 1) were transfected with indicated mirna precursors and

5 Luo, Wen, Wang et al Figure 2. mir-200c and 150 play an important role in endothelial cell (EC) differentiation from human embryonic stem (hes) cells. (A): A panel of five mirnas were upregulated during EC differentiation from hes cells. Total RNA including small RNA were extracted and subjected to real time quantitative polymerase chain reaction (RT-qPCR) analysis to examine mirna expression. Day 0 samples were undifferentiated hes cells and served as negative control. The data presented here are mean6s.e.m. of three independent experiments. *, p <.05. (B D): EC gene expressions were inhibited by mir-200c or 150 inhibitors. Day 6 differentiating hes cells were transfected with mir-200c, mir-150 inhibitor, or mir inhibitor negative control and cultured in endothelial cell growth medium-2 for hours. Total RNA and protein were harvested and subjected to RT-qPCR (B, C) and Western blot (D) analyses, respectively. Bars in the (D) represented the quantitative data of the relative protein levels of CD144 and CD146 from three independent experiments. The data presented here are representative or mean 6 S.E.M. of three independent experiments. *, p <.05. Abbreviation: mirna, micro RNA. cultured in EGM-2 medium for 3 days. Data showed that while mir-200c or 150 over-expressions significantly increased CD144-positive cell numbers, no significant difference was observed in terms of CD144 protein expression levels (the mean fluorescence intensity) within individual ECcommitted cells (CD144-positive cells) (Fig. 3E), indicating that mir-200c or 150 upregulated EC gene expression levels by increasing EC-committing/committed cell number in differentiating hes cells. We further wonder whether mir-200c or 150 play any role in EC proliferation. To achieve this

6 1754 MicroRNAs in Endothelial Cell Differentiation Figure 3. mir-200c and 150 promote endothelial cell (EC) gene expression by increasing EC committing/committed cells from differentiating human embryonic stem (hes) cells. (A C): Over-expression of mir-200c or 150 upregulated EC gene expression. Day 6 differentiating hes cells were transfected with mir-200c, mir-150 precursor, or mir precursor negative control and cultured in endothelial cell growth medium-2 (EGM- 2) for hours. Total RNA and protein were harvested and subjected to real time quantitative polymerase chain reaction (RT-qPCR) (A, B) and Western blot (C) analyses, respectively. Bars in the (C) represent the quantitative data of the relative protein levels of CD144 and CD146 from three independent experiments. (D): mir-200c and 150 increased EC committing cell number (CD146-positive cells) from differentiating hes cells. Day 6 differentiating hes cells were transfected with indicated mirna precursors and cultured in EGM-2 for 3 days. Cells were harvested and subjected to flow cytometry analysis with CD146 antibody. (E): Overexpression of mir-200c and 150 increased fully differentiated EC cell number (CD144-positive cells) but not the EC specific protein expression levels within the fully differentiated cells. Sorted CD146 cells (freshly isolated or passage one) were transfected with indicated mirna precursors, and cultured in EGM-2 medium for 3 days. Cells were harvested and subjected to flow cytometry analysis with CD144 antibody. The numbers and mean fluorescence intensity (MFI) of CD144-positive cells were presented here, respectively. (F): mir-200c and 150 play no role in EC proliferation of fully differentiated ECs (CD146-positive cells, passage 5 10) or HuVECs were transfected with indicated mirna precursors, and cultured in EGM-2 for 3 days. Cell numbers were counted and presented here. The data presented here are representative or mean6s.e.m. of three independent experiments. *, p <.05. Abbreviations: HuVEC, human umbilical vein endothelial cells; mirna, microrna. subjective, fully differentiated ECs (CD146-positive cells, passage 5 10) or HuVECs were transfected with indicated mirna precursors and cultured in EGM-2 medium for 3 days. We found mir-200c and 150 played no role in EC proliferation (Fig. 3F). Finally, our data also suggest that mir-200c and 150 play no significant role in EC specific gene expression in the fully differentiated ECs (supporting information Fig. S5). MiR-150 and 200c Express in EC-Committing Cells and Promote Mesodermal Differentiation Among the upregulated mirnas, mir-141, 205, and 1915 have been suggested to play no role in terms of EC gene regulation in our differentiation model (supporting information Fig. S4). Since our screening for microrna expression during EC differentiation and mirna inhibition experiments were mainly done on mixed differentiating cell cultures, therefore, we further wondered if the increased mirnas expressed or upregulated in endothelial-committing/committed cells only or are they also expressed or upregulated in other cells that may codifferentiate with ECs in this culture system. To address such possibilities, the mirna expression patterns in different cells were examined. Data shown in supporting information Figure S6 clearly revealed that while mir-200c and 150 were upregulated/enriched in CD146-positive cells (EC-committing cells) and still displayed a substantial higher

7 Luo, Wen, Wang et al expression level in the fully differentiated ECs and HuVECs, other three mirnas were significantly decreased in CD146- positive cells. On the contrast, while both the expression levels of mir-200c and 150 were significantly decreased in CD146-negative cells, no big difference in terms of the expression levels of the rest three examined mirnas was observed between day 7 differentiating cells and CD146- negative cells, suggesting that these three mirnas may play a role in non-ec differentiation from stem cells. It has been suggested that mesodermal differentiation is an important intermediate step toward endothelial differentiation during development, and our above data revealed that mir-200c or 150 plays an essential role in EC differentiation. It would be interesting to investigate the functional effects of mir150 and mir200c on germ layer marker gene expression. To this aim, the gene expression levels of three germ layers, such as SOX17 (endoderm), Otx2 (ectoderm), and Brachyury/ T (mesoderm), have been examined in the cells overexpressing mir-200c or 150, respectively. We observed that while the mesoderm gene (T) expression was significantly upregulated by mir-200c and 150 over-expression, neither the endoderm (Sox17) nor the ectoderm (Otx2) gene expression levels was dramatically affected by mir-200c and 150 over-expressions (supporting information Fig. S7), indicating that mir-200c or 150 plays a role in promoting mesodermal differentiation. ZEB1/TCF8 Was Identified As a Communal Target Gene of MiR-150 and 200c To further investigate the mechanisms by which mir-150 and 200c regulate endothelial biology, we looked at potential mrna targets. By using several computational algorithmic databases, such as pictar ( miranda ( and microrna target ( Copoeia.com), we found that there are two highly conserved binding sites for mir-200c and one highly conserved binding site for mir-150 in the 3 0 -UTR of ZEB1 mrna (supporting information Fig. S8). Importantly, ZEB1 gene expressions were significantly downregulated as demonstrated by RTqPCR (Fig. 4A) and displayed an excellent inverse correlation with the gene expression levels of both mir-150 and 200c during EC differentiation from hes cells (Fig. 2A). Moreover, ZEB1 gene and protein levels were significantly downregulated by over-expression of either mir-150 or 200c in the differentiating hes cells (Fig. 4B, 4C), suggesting that ZEB1 may function as the communal target gene of mir-150 and 200c. To confirm such a possibility, the 3 0 -UTR of ZEB1/ TCF8 which contained the binding sites of mir-200c and mir-150 was cloned into a luciferase reporter. Data from our mirna reporter assay showed that the activity of luciferase from construct harboring the ZEB1/TCF UTR was significantly repressed by the over-expression of mir-200c or mir- 150 (Fig. 4D). We further revealed that their respective binding site(s) of mir-200c and 150 are required for mir-150 or 200c mediated ZEB UTR reporter activity repression by using site-directed mutagenesis of the predicted individual binding site or all three binding sites for mir-200c and 150 in the ZEB UTR reporter (Fig. 4E, 4F), confirming that ZEB1/TCF8 is a communal target gene of mir-150 and 200c during EC differentiation. ZEB1/TCF8 Is a Potential Transcription Repressor of EC Gene Regulation A previously published study suggests that ZEB1/TCF8 functions as epithelial-to-mesenchymal transition (EMT) activator by acting as a transcription repressor of E-cadherin, the master regulator of EMT [37]. We investigated if ZEB1/TCF8 plays a role in mir-200c and 150 mediated EC differentiation from hes cells. To this end, ZEB1 knockdown in the differentiating hes cells was conducted by using specific ZEB1 sirna pools. We demonstrated that the knockdown of endogenous ZEB1 significantly upregulated EC specific gene expression (Fig. 5A), suggesting that ZEB1/TCF8 downregulation can recapitulate the effects of mir-200c and 150 during EC differentiation from hes cells. To further investigate if ZEB1 can regulate EC gene promoter activity, EC specific gene reporter plasmids (pgl3-cd144/enos/vwf/flk1/flt1-luc) were generated in this study and used in luciferase activity assay. Data shown in Figure 5B revealed that the knockdown of endogenous ZEB1 in differentiating ES cells or CD146-positive cells significantly increased all the EC specific gene promoter activities, indicating that ZEB1 directly repress transcriptional activity of these EC specific genes. Moreover, by using transcription factor searching software, we identified multiple ZEB1 binding sites within the promoter regions of EC-specific genes including CD144 and enos. As shown in supporting information Fig. S9, three ZEB1 binding sites can be found within 2.5-kb of the VE-cadherin/CD144 gene promoter. Therefore, we further hypothesized that these binding sites are required for ZEB1- regulated EC gene repression. To test this hypothesis, by using pgl3-cd144-luc plasmid as a template, we carried out sitedirected mutagenesis to mutate the ZEB1 binding sites within gene promoters. We successfully mutated Z-box, E-box-2, and both of these binding sites but failed to mutate E-Box-1 binding site in the CD144 gene reporter. Data from our luciferase activity assay showed that while binding sites Z-box and E- box-2 mutation in pgl3-cd144-luc partially reduced CD144 gene transcription activity induced by ZEB1 inhibition, combinational mutations of Z-box and E-box-2 resulted in complete lose of their transcriptional activity in response to ZEB1 knockdown (Fig. 5C). This suggested that ZEB1 regulates CD144 gene expression through these two binding sites within its promoter region. Finally, by using CHIP assay we demonstrated that ZEB1 can directly bind to the promoters of CD144 gene, and such binding was significantly inhibited by the knockdown of endogenous ZEB1 gene level (Fig. 5D). Taken together, these findings strongly suggest that ZEB1 transcriptionally repress EC specific gene expression during EC differentiation from hes cells. MiR-200c and 150 Regulate EC Gene Expression During EC Differentiation from hes Cells by Repressing Transcriptional Repressor ZEB1/TCF8 We have demonstrated that ZEB1/TCF8 is a communal target gene for mir-200c and 150 during EC differentiation from hes cells and is a potential EC gene transcriptional repressor. We further hypothesized that knockdown of endogenous ZEB1 in differentiating hes cells can rescue the inhibitory effects of mir-200c or 150 inhibitions in EC gene expression. To test this, control mirna inhibitor or mir-200c/150 inhibitor and control random sirna or ZEB1 sirnas (30 nm) were cotransfected into day 6 differentiating hes cells. Data shown in Figure 6 revealed that while mir-200c or 150 inhibition (second columns) or knockdown of endogenous ZEB1/TCF8 (third columns) alone in the differentiating hes cells were significantly downregulated or upregulated various EC specific gene expression, respectively, knockdown of endogenous ZEB1/ TCF8 gene expression significantly rescued EC specific gene repression induced by mir-200c or 150 inhibition (fourth columns), suggesting that mir-200c and 150 regulate EC gene expression during EC differentiation from hes cells through targeting and repressing the transcription repressor ZEB1/TCF8.

8 1756 MicroRNAs in Endothelial Cell Differentiation Figure 4. mir-200c and 150 target transcription factor ZEB1 mrna at 3 0 -UTR. (A): ZEB1 was downregulated during endothelial cell differentiation from human embryonic stem (hes) cells. The data presented here are mean 6 S.E.M. of three independent experiments. *, p <.05. (B, C): mir-200c and -150 over-expressions resulted in ZEB1 gene repression. Day 3 differentiating hes cells were transfected with mir-200c, mir- 150 precursor, or mir precursor negative control, and cultured in EGM-2 medium for hours. Total RNA and protein were harvested and subjected to quantitative real time polymerase chain reaction (B) and Western blot (C) analyses, respectively. (D): mir-200c or -150 overexpression repressed pmir-luc-zeb1 luciferase activity UTR of ZEB1/TCF8 was cloned into a luciferase reporter. Then luciferase activity assay was carried out as described in the main text on CD1461 cells or day 6 differentiating hes cells after overexpressing mir150 and 200c at 48 hours post-transfection. (E): Mutation of mir-200c binding site one (mir-200c BS1), binding site two (mir-200c BS2), or mir-150 binding site (mir-150 BS) abolishes mir-200c- or mir-150-mediated repression of pmir-luc-zeb1 luciferase activity, respectively. (F): Mutation of all three mir-200c and 150 binding sites within 3 0 -UTR of ZEB1 (pmir-zeb1 3 mutations) completely abolished mir-200c and mir-150-mediated luciferase activity. The data presented here are representative or mean6s.e.m. of three independent experiments. *, p <.05. Abbreviations: mirna, microrna; ZEB1, zinc finger E-box-binding homeobox 1. MiR-200c and 150 are Involved in Blood Vessel Formation or Vasculogenesis In Vivo Since it was previously shown in vitro that mir-200c and 150 modulate the expression of EC markers in differentiating hes cells, it will be interesting to discern if this phenomenon also occurs in vivo. Chicken embryo models are valuable tools for testing in vivo gene functions of SMC differentiation regulators identified from our in vitro SMC differentiation model in the smooth muscle development and arteriogenesis as demonstrated in our previous studies [38,39]. Moreover, chick embryo is easy to work with and readily available. Therefore, we recently established a procedure by using developing chick embryos to investigate the functions of mir- NAs in chicken blood vessel formation or vasculogenesis.

9 Luo, Wen, Wang et al Figure 5. ZEB1 is a potential transcription repressor of endothelial cell (EC) gene expression. (A): Knockdown of ZEB1 upregulated EC gene expression. Day 6 differentiating human embryonic stem cells were transfected with ZEB1 sirna or control sirna and cultured in endothelial cell growth medium-2 for another 48 hours. Total RNAs were harvested and subjected to real time quantitative polymerase chain reaction analyses. (B): ZEB1 knockdown upregulated EC gene promoter activity. CD1461 cells were cotransfected with respective EC gene luciferase reporter plasmids (0.15 mg per cells) together with control sirna or ZEB1 sirna (20 nm). pgl3-renilla (0.025 mg per cells) was included as luciferase plasmid control. Luciferase and Renilla activity assays were detected 48 hours after transfection. (C): ZEB1 binding sites within CD144 gene promoter are required for ZEB1-mediated CD144 gene promoter activity. (D): ZEB1 bound directly to the promoter regions of CD144 gene. ChIP assays were performed using antibodies against ZEB1 or normal rabbit IgG, respectively, as described in supporting information data. PCR amplifications of the non Z-box region were included as additional control for specific promoter DNA enrichment. The data presented here are mean 6 S.E.M. of three independent experiments. *, p <.05. Abbreviations: ChIP, Chromatin immunoprecipitation; sirna, small interfering RNA; ZEB1, zinc finger E-box-binding homeobox 1. Data shown in Figure 7A revealed that treatment with transfection complexes containing mir-200c inhibitor resulted in a clear impairment in terms of blood vessels formation or vasculogenesis as compared to control treated chick embryos. Similar phenomena were also observed by using mir-150 inhibitor (Fig. 7A). Importantly, RT-qPCR analyses showed that mir-200c and 150 expression levels in the chick choroallantoic membrane transfected with mir-200c or 150 inhibitors were significantly lower than that of control mirna inhibitor transfected choroallantoic membrane (Fig. 7B). On the other hand, the ZEB1/TCF8 gene expression levels were significantly increased in the chick choroallantoic membrane treated with mir-200c or 150 inhibitors (Fig. 7B). These observations further confirmed the knockdown efficiency of mir-200c and 150 in the choroallantoic membrane and the importance of mir-200c/150-zeb1 axis in chick embryonic vasculogenesis. To further explore its importance of mir-200c in mammalian vasculogenesis and/or the functional impacts of the combinatory manipulation of mir-200c and ZEB1 signaling in such process, CD146-positive cells (passage 1, EC-committing ECs) were labeled with PKH26, and mixed with Matrigel contained VEGF165 and siporttm NeoFXTM transfection reagent. Such cell-matrigel mixtures were aliquot into three tubes (I III) and mixed with control mirna inhibitor/sirna (I), mir-200c inhibitor/control sirna (II), or mir-200c inhibitor/zeb1 sirna (III), respectively. The resultants were subcutaneously injected into SCID mice for 2 weeks. Data from immunofluorescence staining analyses showed that while the percentage of the CD144- (Fig. 7C) and CD31- (supporting information Fig. S10) positive cells within Matrigel implants was significantly inhibited by mir-200c inhibition, such inhibitory effect was rescued by addition of ZEB1 sirna, indicating that mir-200c regulates in vivo vasculogenesis or EC differentiation through modulating ZEB signaling. As expected, RT-qPCR analyses showed that the endogenous mir-200c and ZEB1 levels were successful knockdown within Matrigel implants (Fig. 7D). Taken together, these striking observations strongly imply that the EC differentiation regulator or signaling pathway mir-200c/ 150-ZEB1 axis, plays an important role in in vivo vasculogenesis. DISCUSSION In this study, we have established a procedure to produce large numbers of ECs from hes cells ( ECs from

10 1758 MicroRNAs in Endothelial Cell Differentiation Figure 6. Knockdown of endogenous ZEB1 gene expression rescues the inhibitory effects of mir-200c or 150 inhibitions in endothelial cell gene expression. Day 6 differentiating human embryonic stem cells were transfected with ZEB1 sirna or control sirna and control mirna inhibitor, mir-200c inhibitor (A) or mir-150 inhibitor (B), then cultured in endothelial cell growth medium-2 for another 48 hours. Total RNAs were harvested and subjected to quantitative real time polymerase chain reaction analyses. The data presented here are mean 6 S.E.M. of three independent experiments. *, p <.05 (vs. control treatment); #, p <.05 (mir-200c and ZEB1 sirna vs. mir-200c or ZEB1 sirna alone).abbreviations: mirna, micro RNA; sirna, small interfering RNA; ZEB1, zinc finger E-box-binding homeobox hes cells) which may potentially be used as cellular therapy following vascular injury. We first isolated CD146- positive from the heterogeneous differentiating hes cells, which can be further differentiated into functional ECs with high purity. Second, we have demonstrated for the first time that mir-200c and 150 play an important role in EC differentiation from hes cells. Third, we have successful identified ZEB1/TCF8 as a communal target gene of both mir-200c and 150 during EC differentiation and clearly defined that ZEB1/TCF8 is a potential transcriptional repressor of EC gene regulation. Finally, but most importantly, we have further demonstrated that mir-200c and 150-ZEB1 signaling axis plays a critical role in chick embryonic vascularogenesis by using choroallantoic membrane (CAM) assay in developing chick embryos. These findings strongly suggest that CD146 1 cells derived from hes cells may be used to generate a large number of ECs, which can provide a promising cell source for cell-based vascular engineering and repair of injured vessels. Moreover, our findings also provide new insights into the molecular mechanisms involved EC differentiation from hes cells and better understanding of the cardiovascular system development. MiR-200c/150 and Cell Differentiation In mammalians, the mature mir-150 is highly expressed in the lymph nodes, spleen, brain, heart, thymus, and detectable in the liver but not expressed in lung, muscle, bone marrow, kidney, and adipose tissue [40]. Recent evidence suggests that mir-150 can regulate embryonic development [41] and hematopoietic cell lineage development/differentiation such as early B cell [40,42], natural killer (NK) and invariant NK T cell [43], and T cell [44], by targeting c-myb or Notch3. On the other hand, mir-200c, a member of the mir-200 family

11 Luo, Wen, Wang et al Figure 7. Functional impacts of mir-200c or 150 on chick embryonic blood vessel formation and in vivo endothelial differentiation. (A): In vivo inhibition of mir-200c or 150 impaired chick embryonic blood vessel formation. The right-hand side of HH10 staged chick embryos were treated with transfection complexes containing anti-mir negative control (upper), mir-200c inhibitor (middle), or mir-150 inhibitor (bottom) as described in main text. The relative area of blood vessels formed in these chick embryos were analyzed later at HH20 stage. Left, the representative images of chick embryo (HH20) administered with negative control mir inhibitor, mir-200c inhibitor, or mir-150 inhibitor, respectively. Right, the enlarged image of the selected area in the left images. Far right, quantitative analysis of the relative blood vessel area from both sides of the chick embryo using Image J plus. Each column represent mean 6 SEM of three independent experiments (n 5 3). *Relative area of blood vessels with mir-200c or mir-150 inhibition (R) was significantly different from nontreated area (L) at HH20; p <.05. (R) right, (L) left. (B): mir-200c and mir-150 inhibitors administration result in chick embryonic blood vessel developmental retardation through derepression of ZEB1. Total RNA including small RNA were harvested from chick CAM (HH20) and subjected to real time quantitative polymerase chain reaction (RT-qPCR) analysis. *, p <.05 (vs. control mirna treatment). (C): ZEB1 knockdown rescues the inhibitory effects of mir-200c inhibition in in vivo vasculogenesis. Matrigel plugs implanted with sorted CD146-positive cells transfected with sirna and mirna inhibitors as indicated (groups I to III) were harvested, sectioned, and subjected to immunofluorescence (C) staining analyses using mouse anti human CD144 antibody (ab166715). Representative images (left panel) and quantitative data (right bar graphs) of the percentage of CD144-positive cells were presented here, respectively. Note: Cells with red fluorescence signal indicate PKH26-labeled cells (implanted cells) in panel (C). The percentage of PKH26-labeled CD144-positive cells per field were examined by two well-trained independent investigators blinded to the treatments, from four random high power fields (3200) in each section, three sections from each implant, and three implants for each group. *, p <.05 (group II versus I); **, p <.05 (group III versus II). (D): Expression levels of mir-200c and ZEB1 within Matrigel implants. Total RNA samples were extracted from partial Matrigel implants and subjected to real-time PCR analysis. The data presented here were mean 6 S.E.M. of three mice. *, p <.05 (group II versus I); #, p <.05 (group III versus II). Abbreviations: DAPI, 4 0,6-diamidino-2-phenylindole; mirna, microrna; sirna, small interfering RNA; ZEB1, zinc finger E-box-binding homeobox 1. found clustered in chromosome 12, is highly expressed in epithelial tissues [45] and has been implied to play a role in EMT in response to transforming growth factor beta or to ectopic expression of the protein tyrosine phosphatase Pez by targeting E-cadherin transcriptional repressors ZEB1/TCF8 [46]. EMT facilitates tissue remodeling during embryonic development and has been highlighted in mesoderm formation/differentiation to diverse tissues or organs during embryogenesis [47], indicating that mir-200c might play a role in cell differentiation. Moreover, recent evidence suggests

12 1760 MicroRNAs in Endothelial Cell Differentiation that mir-200c also play a role in reactive oxygen speciesinduced EC apoptosis and senescence through targeting ZEB1/TCF8 [48], and ZEB1/TCF8 is a negative regulator in tumor angiogenesis [49]. However, there is no evidence in the literature to suggest if mir-200c or 150 play a role in EC differentiation and vascular system development. By using our EC differentiation system and mirna array technique, and using mirna gain/lose-of function analyses, we confirm a critical role of mir-150 and 2200c in EC differentiation from hes cells. We provide further evidence on the function of mir-150 and 200c in embryonic EC differentiation and/or vascular development by inhibiting their expression in developing chick embryos. These observations clearly implied that mir-150 and mir-200c do play an important role in EC differentiation from hes cells and chick embryonic vasculogenesis. However, one should be cautious when considering the importance of mir-150 in vasculogenesis since mir-150- deficient mice were viable, fertile, and morphologically normal [42], although such discrepancies could be attributed to the compensatory effects of other vasculogenesis regulators/ molecules such as mir-200c in the mir-150 knockout mice, different species and animal models used in the studies (mir- 150 knockout mice vs. Chick CAM assay), and the fact that the signaling networks by which mir-150 mediates in vivo vasculogenesis in mice is much more complicated than that in in vitro cell culture system. Moreover, ZEB1/TCF8 has been identified as the communal target of mir-200c and 150 in EC differentiation and vasculogenesis in this study, and it has been implicated in tumor angiogenesis [49], further studies using the double knockout mice, such as mir-200c/150 or mir-150/zeb1, would be warranted to fully understand the functional importance of mir-200c or 150 in in vivo vasculogenesis or blood vessel formation. Also it would be very interesting to see if mir-150 deficiency has any functional impacts in postnatal vasculogenesis or angiogenesis under physiological and/or pathological conditions in these mice. It is noteworthy that along with mir-200c and 150, three other mirnas (mir-141, 205, and 1915) were also significantly upregulated in our EC differentiation system. Inhibition studies were carried out with these three mirna inhibitors that inhibit endogenous mirnas to examine their potential functions in EC differentiation from hes cells. Our data in inhibition studies showed that although all three of the selected mirnas were inhibited by their respective inhibitor in the differentiating hes cells, none of them downregulated the mrna expression of specific EC markers as compared to control levels (supporting information Fig. S4), indicating that these three mirnas play no role in terms of EC gene regulation in our differentiation model. Additionally, our data also revealed that mir-200c and 150 were specifically upregulated in EC-committing cells, while other three examined mirnas were significantly decreased in such cells (supporting information Fig. S6). Interestingly, we also observed that mir- 200c and 150 only significantly promote mesoderm gene expression (supporting information Fig. S7), suggesting that mir-200c or 150 play a cell-autonomous role in mediating EC differentiation from stem cells. ZEB1/TCF8 As a Potential EC Gene Transcription Repressor? One of our novel findings in this study is that we have successfully indentified ZEB1/TCF8 as a communal target gene of the mir-200c and 150 in EC specific gene regulation. mrna levels of ZEB1/TCF8 were significantly decreased as compared to day 0 control (undifferentiated hes cells). Moreover in mirna over-expression studies, it was shown that mrna expression of ZEB1/TCF8 were decreased by overexpression either mir-200c or 150. These observations demonstrated that ZEB1/TCF8 mrna expression levels were reciprocally regulated by mir-150 and 200c, which is consistent with the presence of binding sites for both mir200c and mir-150 in the 3 0 -UTR of ZEB1/TCF8 (supporting information Fig. S8). As expected, upon over-expressing mir-200c or 150 in the differentiating hes cells and/or CD1461 cells, luciferase activity of 3 0 -UTR of ZEB1/TCF8 was significantly reduced. Importantly, binding sites mutations were carried out to further elucidate if the respective bindings are required for mir-200c or 150-regulated ZEB1 gene repression, and luciferase activity assay with respective mutant firmly confirmed their importance of each binding site in ZEB1 gene repression. This demonstrates that both mir-150 and 200c directly target ZEB1/TCF8 and subsequently leading to the differentiation of mature ECs from undifferentiated hes cells. Another major finding in this study is that we have provided strong evidence to support its transcriptional repressor s role of ZEB1/TCF8 in EC gene regulation during EC differentiation from hes cells. Data from loss-of-function gene studies using specific ZEB1 sirna pool showed that the knockdown of endogenous ZEB1 gene expression significantly increased EC-specific gene expression levels and promoter activity. Importantly, we further demonstrated that ZEB1 binding sites (Z-box and E-box-2) are required for ECspecific gene repression regulated by ZEB1 by using sitedirected mutagenesis technique. Finally, by using CHIP assay we also confirmed that ZEB1 regulates EC gene expression through direct binding to the promoter region of EC genes. All these data strongly suggest that ZEB1 is a potential transcriptional repressor of EC gene regulation, which is consistent with previous findings that ZEB1 usually acts as a transcription repressor [50]. ZEB1 is expressed in proliferating mesenchymal and neural progenitors, and mutation of the ZEB1 gene in mice leads to severe T-cell deficiency of the thymus and skeletal defects of various lineages [51]. It also causes mesenchymal epithelial transition in gene expression and diminished proliferation in various progenitor cells at sites of developmental defects in mouse embryos [52]. It also appears that ZEB1 is expressed in less differentiated cell types like ES cells [53,54]. Moreover, Wellner et al. [54] also discovered that during murine ES cell differentiation, levels of mir-200c was upregulated while stem cell factors such as KLF4 and Sox2 were decreased. In another study, deep sequencing analysis also indicated that expression of mir- 200c was increased upon differentiation of hes cells [55]. It has been reported that in tumor cells certain mirnas including mir-200c were repressed by ZEB1/TCF8 over-expression [56], and reciprocally, mir-200c and other mir-200 family members repressed ZEB1 expression by directly targeting its 3 0 -UTR [57]. Therefore, it has been proposed that ZEB1 and mir-200 family members are interconnected through doublenegative feedback loop, and such interconnection plays an important role in tumor cell invasion and metastasis [46,53]. Taken together, the findings from this study and previous studies strongly suggest that as hes cell differentiation towards the EC lineage, upregulated mir-200c or 150 can simultaneously repress ZEB1/TCF8 gene expression by directly targeting its 3 0 -UTR, which in turn de-repress EC gene expression and result in EC differentiation and embryonic vascular formation/development. Moreover, an increase of mir-200c or 150 expression may contribute to the decline or repression of other cell lineage specification processes such as skeletal cell lineages and hematopoietic cell lineage specifications. Subsequently, this promotes EC differentiation from stem cells and vasculogenesis.