1951
T
he development of the vascular network is a highlyorga-nized process composed of correct spatial connections and regulated gene expression, which is essential for the integrity of the endothelial monolayer in blood vessel homeo-stasis. These events involve several genes including growth factors, cytokines, cell–cell connection factors, and tion factors. To date, a few endothelial cell–specific transcrip-tion factors have been identified, and it has been suggested that a combination of several transcription factors, such as Hox, Forkhead box protein O, Ets, GATA family members,1 vascular endothelial zinc finger 1,2 stem cell leukemia/T-cell acute lymphoblastic leukemia 1,3 and hairy/enhancer of split related with YRPW motif protein 1 and 2,4 together with other nonendothelial-specific transcription factors, control gene expression in endothelial cells. However, it remains unclear how specific transcription factors serve as master switches in controlling transcriptional regulation and coordinating differ-ent cell–cell connection molecules to maintain endothelial integrity.
Related transcriptional enhancer factor-1 (RTEF-1) is 1 of the 4 vertebrate members of the transcriptional enhancer
factor-1 multigene family. The transcriptional enhancer factor-1 transcriptional system plays important roles in a wide variety of physiological and pathological conditions. RTEF-1 deficiency causes defects in the production of trophoblast stem cells, trophectoderm, and blastocoel cavities and, conse-quently, preimplantation lethality.5 RTEF-1 is targeted at the promoters of many genes expressed in cardiac,6 skeletal, and smooth muscle cells.7 Recent lines of evidence have impli-cated RTEF-1 in the regulation of angiogenesis under hypoxia via stimulating vascular endothelial growth factor (VEGF) expression.8 However, the role of RTEF-1 in vascular cells is still unclear.
In this study on endothelial network formation in vitro and in vivo, we describe a novel endothelial function of RTEF-1 as a vascular structure modulator. Endothelial cell prolifera-tion, migraprolifera-tion, and the formation of networks occur during the maturation of primitive tubes into functional blood vessels. Extending our finding that RTEF-1 enhanced endothelial cell proliferation,8 we observed that RTEF-1 also enabled endothelial cells to connect, aggregate and dramati-cally alter the pattern of endothelial network formation.
Received on: November 17, 2011; final version accepted on: May 7, 2012.
From the Institute of Molecular Medicine, Peking University, Beijing, China (X.A., Y.J., J.L.); and Cardiovascular Institute, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA (X.A., Y.J., M.J.P., J.W., A.M.-B., X.S., B.L.C., P.H., M.X., H.S.D., J.L.).
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.112.250159/-/DC1. Correspondence to Jian Li, MD, PhD, Division of Cardiovascular Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, 3 Blackfan Circle, Boston, MA 02115. E-mail JLi@BIDMC.Harvard.edu
© 2012 American Heart Association, Inc.
Arterioscler Thromb Vasc Biol is available at http://atvb.ahajournals.org DOI: 10.1161/ATVBAHA.112.250159
Arterioscler Thromb Vasc Biol 1079-5642 10.1161/ATVBAHA.112.250159 ATV201446 00 00 0 0
© 2012 American Heart Association, Inc.
28 August 2012
An et al
The Role of RTEF-1 in Endothelial Cell Connections
Objective—Capillary network formation represents a specialized endothelial cell function and is a prerequisite to establish
a continuous vessel lumen. Formation of endothelial cell connections that form the vascular structure is regulated, at least in part, at the transcriptional level. We report here that related transcription enhancer factor-1 (RTEF-1) plays an important role in vascular structure formation.
Methods and Results—Knockdown of RTEF-1 by small interfering RNA or blockage of RTEF-1 function by the
transcription enhancer activators domain decreased endothelial connections in a Matrigel assay, whereas overexpression of RTEF-1 in endothelial cells resulted in a significant increase in cell connections and aggregation. In a model of oxygen-induced retinopathy, endothelial-specific RTEF-1 overexpressing mice had enhanced angiogenic sprouting and vascular structure remodeling, resulting in the formation of a denser and more highly interconnected superficial capillary plexus. Mechanistic studies revealed that RTEF-1 induced the expression of functional gap junction proteins including connexin 43, connexin 40, and connexin 37. Blocking connexin 43 function inhibited RTEF-1–induced endothelial cell connections and aggregation.
Conclusion—These findings provide novel insights into the transcriptional control of endothelial function in the coordination
of cell–cell connections. (Arterioscler Thromb Vasc Biol. 2012;32:1951-1959.)
Key Words: related transcription enhancer factor-1 n endothelial cells n gap junctions n connexin 43
n cell–cell connections
Endothelial Cells Require Related Transcription Enhancer
Factor-1 for Cell–Cell Connections Through the Induction
of Gap Junction Proteins
Xiaojin An, Yi Jin, Melissa J. Philbrick, Jiaping Wu, Angela Messmer-Blust, Xiaoxiao Song,
Brittany L. Cully, Ping He, Ming Xu, Heather S. Duffy, Jian Li
We further found that gap junction proteins, specifically connexin (Cx) 37, Cx40, and Cx43, were significantly regulated by RTEF-1 and acted as adhesion sites that enhanced endothelial cellular connections and aggregation, as well as promoted the timely formation of competent vascular networks. We selected a mouse retina model for our in vivo studies because of the retinal vasculature developing shortly after birth in a highly stereotypic manner,9 and RTEF-1 appeared as an essential gene for retinal vascular development by enhancing angiogenic sprouting and promoting the formation of a denser primary capillary network in an oxygen-induced retinopathy (OIR) model. These observations indicate that RTEF-1 coordinated endothelial cell-to-cell connections and is essential for the assembly of vascular networks in vitro and in vivo.
Materials and Methods
Cell Culture
Human umbilical vein endothelial cells (Lonza, Allendale, NJ) were
cultured in endothelial basal medium-2 (Lonza). Human microvas-cular endothelial cells-1 (HMEC-1, Center for Disease Control and Prevention, Atlanta, GA) were cultured in MCDB-131 (Invitrogen, Carlsbad, CA). HMEC-1 is an immortalized HMEC line that retains the morphological, phenotypic, and functional characteristics of nor-mal HMEC.10 Human embryonic kidney 293 and 293T were cultured
in DMEM (Invitrogen).
Retroviral Transduction and Small Interfering RNA Transfection
The coding sequence of RTEF-1 (NM_003213) was subcloned from a PXJ40/RTEF-1 construct (a gift from Dr Alexandre Stewart, University of Ottawa) into the pBMN-GFP vector (Orbigen, San Diego, CA) to produce pBMN-GFP-RTEF-1. 293T cells were transfected with pBMN-GFP (or pBMN-GFP-RTEF-1), pMD-VSVG, pJK3, and pCMV-tat using Polyethylenimine (Polysciences, Warrington, PA) as previously reported11 to a package virus. The
virus-containing medium was transduced to HMEC-1 and selected with puromycin (250 ng/ml). RTEF-1 small interfering RNA (siRNA) was synthesized by Qiagen (Valencia, CA), and short hairpin RNA (shRNA) was purchased from SABiosciences (Frederick, MD). Cx43 siRNA was synthesized by Genpharma Inc (Shanghai, China). siRNA not targeted at any human gene was used as a negative con-trol (sequence: 5′-UUC UCC GAA CGU GUC ACG UTT-3′, 5′-ACG UGA CAC GUU CGG AGA ATT-3′). HMEC-1 were transfected with siRNA at a final concentration of 50 nmol/L with Lipofectamine 2000 (Invitrogen). The target sequences of siRNAs are as follows: RTEF-1 siRNA1 5′-AAACCCAAGATGCTGTGTATT-3′ RTEF-1 siRNA2 5′-CAGGTGCTGGCTCGTCGCAAA-3′ RTEF-1 shRNA1 5′-CCCATGATGTGAAGCCTTTCT-3′ RTEF-1 shRNA2 5′-GTGGACATCCGCCAAATCTAT-3′ Cx43 siRNA1 5′-AACAAGCAAGCAAGTGAGCAA-3′ Cx43 siRNA2 5′-AAAGGAAGAGAAACTGAACAA-3′
Matrigel Analysis
Growth factor–reduced BD Matrigel Matrix (BD Biosciences, San Jose, CA) was coated on a prechilled 12-well culture plate on ice. After Matrigel solidification for 30 minutes in 37°C incubator, HMEC-1 were plated 105 cells/well. The extent of network formation
during various time points was photographed using a TS100 Nikon microscope (Nikon, Melville, NY). The network was quantified with Image J software to calculate the aggregation area and branch number.
RNA Analysis
Total RNA was extracted using TRIzol Reagent (Invitrogen), and cDNA was synthesized by reverse transcription. RNA levels were measured by quantitative polymerase chain reaction. All quantitative polymerase chain reaction data are represented in the results as a ratio relative to the housekeeping gene, 36b4. Primers used are as follows: N-cadherin F: CAACTTGCCAGAAAACTCCAGG R: ATGAAACCGGGCTATCTGCTC VE-cadherin F: GACCGGGAGAATATCTCAGAGT R: CATTGAACAACCGATGCGTGA β-catenin F: TACCTCCCAAGTCCTGTATGAG R: TGAGCAGCATCAAACTGTGTAG Claudin 5 F: CTCTGCTGGTTCGCCAACAT R: CAGCTCGTACTTCTGCGACA Occludin F: CCCTTTTAGGAGGTAGTGTAGGC R: CCGTAGCCATAGCCATAACCA ZO-1 F: AGTCCCTTACCTTTCGCCTGA R: TCTCTTAGCATTATGTGAGCTGC ZO-2 F: CAGCGATCAACTTAGGGACAAT R: AGATGCCCCAGGAGTTTCATTA Cx37 F: GACCAGGTCCGAGAGCACT R: CCGTGTTACACTCGAAATCTGA Cx40 F: TCCTGGAGGAAGTACACAAGC R: ATCACACCGGAAATCAGCCTG Cx43 F: TCAAGCCTACTCAACTGCTGG R: TGTTACAACGAAAGGCAGACTG RTEF-1 F: CCACGAAGGTCTGCTCTTTC R: CTCACTGGCTGACACCTCAA 36b4 F: GGCTCCAAGCAGATGCAGCAG R: CCTGATAGCCTTGCGCATGG Western Blot Analysis
Tissue or cells were lysed in radioimmunoprecipitation assay buffer (Boston BioProducts Inc, Ashland, MA) containing a cocktail of pro-tease inhibitors (Roche, Nutley, NJ). Cells plated on Matrigel were recovered with BD Cell Recovery Solution (BD Biosciences, San Diego, CA) before being lysed. Samples were subjected to SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Millipore, Lincoln Park, NJ). After being blocked for 1 hour in Tris-buffered saline-Tween 20 with 5% nonfat milk, the polyvinylidene difluoride membranes were incubated with the indicated antibodies: anti–RTEF-1 (synthesized by Genemed, South San Francisco, CA), anti–connexin 37, anti–connexin 40, anti–connexin 43 (Invitrogen), anti-actin, anti–β-catenin, anti–vascular endothelial (VE)-cadherin (Santa Cruz Biotechnology, Santa Cruz, CA), and anti-vinculin (Sigma, St. Louis, MO) followed by incubation with horseradish peroxidase anti-mouse or anti-rabbit IgG (Calbiochem, San Diego, CA). The blots were detected with the ECL detection system according to the instructions of the manufacturer (Thermo Scientific, Barrington, IL).
Luciferase Assay for Promoter Activity and Chromatin Immunoprecipitation
A 1200-bp fragment of the human Cx43 promoter (−978 to +215)12
was obtained by polymerase chain reaction from human genomic DNA. The Cx43 promoter was cloned into a luciferase reporter vector to obtain pGL3/Cx43 promoter. Then, 200 ng/well of Cx43 promoter construct was transfected into HMEC-1/RTEF-1 and HMEC-1/control in a 24-well plate using Lipofectamine 2000 (Invitrogen). After 24-hour transfection, luciferase activity was determined using the Dual-Luciferase assay system (Promega, Fitchburg, WI). Chromatin immunoprecipitation was performed with the ChIP-IT Express Kit (Active Motif, Carlsbad, CA) in accordance with the manufac-turer’s instructions. The Cx43 and actin promoters were amplified with the primer pairs 5′-TTGCGGTGAGCAGAGATGG-3′ and 5′-GGCCTTCCTATTCCTAACACTT-3′, and 5′-TGCACTGTGCGG CGAAGC-3′ and 5′-TCGAGCCATAAAAGGCAA-3′, respectively. The Cx43 primers were designed to flank a putative specificity pro-tein 1 (SP1) site.
Immunofluorescence Assay
HMEC-1 (control and RTEF-1 overexpression [o/e]) were fixed with 4% paraformaldehyde and incubated with anti–connexin 43 (Sigma; Invitrogen) or anti–VE-Cadherin (Santa Cruz Biotechnology) antibody overnight at 4°C, followed by incubation with Texas
red-conjugated or fluorescein isothiocyanate–conjugated (Santa Cruz Biotechnology) secondary antibody for 1 hour. Slides were mounted using mounting medium with 4′,6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA) and imaged on a Leica fluorescent microscope (Leica, Allendale, NJ) or a Zeiss LSM510 confocal microscope (Zeiss, Thornwood, NY).
Scrape Loading and Dye Transfer Assay
The scrape-loading technique was used to evaluate the degree of gap junctional intercellular communication.13,14 In brief, confluent
HMEC-1 (control and RTEF-1 o/e) were bathed in phosphate-buff-ered saline containing 0.5% Alexa Fluor 350 hydrazide (Invitrogen) and 1% Texas Red (Invitrogen). Incisions on the culture surface were made with a razor blade, allowing dyes to enter the scraped HMEC-1. After 5 minutes of incubation, the cultures were rinsed 3× with PBS, fixed in 4% paraformaldehyde (Sigma), and pho-tographed. The extent of Alexa Fluor 350 hydrazide showed the functional gap junction, and Texas Red indicated dead cells. Dye transfer calculated using Image J software was obtained by deter-mining the maximal perpendicular distance from the incision at which dye could be clearly detected.
Generation of Endothelial-Specific RTEF-1 Transgenic Models
An endothelial-specific inducible transgenic mouse model (tTA+/
RTEF-1+) was generated in our lab based on a strategy previously
reported.15 RTEF-1 cDNA (NM_003213) was subcloned into a
bidi-rectional pBI-G Tet Vector (Clontech, Mountain View, CA) sensitive to tTA. Flanked by 2 minimal cytomegalovirus promoters, tetracy-cline response elements controls LacZ as a reporter gene as well as the RTEF-1 gene. The tetracycline response elements-RTEF-1/LacZ construct was microinjected into fertilized mouse eggs. The endo-thelial-specific expression transgenic mouse, VE-Cadherin:RTEF-1, was obtained by crossing the endothelial-specific driver line (VE-Cadherin:tTA) (from Dr Laura Benjamin, BIDMC)16 and
responder line (tetracycline response elements-RTEF-1/LacZ). Double transgenic mice were genotyped by polymerase chain reac-tion with primers specific for tTA and RTEF-1. Littermates not inher-iting both transgenes were used as controls. Transgene expression of RTEF-1/LacZ was repressed by dietary doxycycline (200 mg/kg food) and induced on a normal diet. Control mice were treated similarly. All animal experiments were approved by the Institutional Animal Care and Use Committee of Beth Israel Deaconess Medical Center.
Postnatal Retinal Vascularization and OIR
Five- to 17-day-old pups were used to assess the effect of RTEF-1 signaling. OIR was produced following the method developed by Smith et al.9,17 Briefly, at the postnatal day 7 (P7), pups were exposed
to constant hyperoxia (75% O2) for 5 days and then returned to nor-moxia for 5 days. During the exposure, mice were given free access to food and water and were housed in plexiglass chambers in which the oxygen level was controlled by a feedback device (OxyCycler; Reming Bioinstruments, Redfield, NY).
Retina Staining and Quantification of Retinal Angiogenesis
Mouse pups were euthanized on P8 and P17.9 Eyes were enucleated
and retinas were dissected, fixed overnight with 4% paraformaldehyde, immunostained with fluorescein isothiocyanate–labeled Griffonia
sim-plicifolia, lectin I (Invitrogen), and Cx43 (Sigma; Invitrogen) antibodies, and flat mounted. Retina staining and quantification of retinal angiogen-esis were based on previous reports.9,18 In the OIR model, the avascular
area was quantified using Image J software. For each eye, the avascular region was selected and quantified in relation to the area size. Vascular structure morphology from 5 different X5 images taken at the vessel branch area was quantified using IDL software by converting the image to black and white followed by manual threshold adjustment to highlight the Lectin-positive area, and the ratio of white pixels was then quantified.
Statistics
The results are expressed as mean±SEM or SD based on 3 indepen-dent experiments. Statistical analysis used ANOVA and Stuindepen-dent t test (2-tailed). P<0.05 was considered statistically significant.
Results
RTEF-1 Promotes Endothelial Cell Connections and Aggregation In Vitro
An increase in capillary network formation is one of the major parameters used to define angiogenic factors. In the Matrigel assay, HMEC-1 cells in which endogenous RTEF-1 was knocked down by siRNA (Figure 1A) exhibited significant attenuation of endothelial cell connections, whereas wild-type HMEC-1 grew as elongated cells with cellular contacts to form a normal endothelial network (Figure 1B). HMEC-1 trans-fected with negative control siRNA demonstrated a similar pattern as wild-type HMEC-1. To further examine the effect of RTEF-1 on cell–cell connections, HMEC-1 cell lines with con-structs containing full-length RTEF-1 (RTEF-1 o/e) or only the DNA binding domain (transcriptional enhancer activators) of RTEF-1 (RTEF-1 dominant negative) were generated (Figure 1C). The Matrigel assay revealed that HMEC-1 cells with forced overexpression of RTEF-1 (RTEF-1 o/e) established an assembly of endothelial networks with increased cell aggrega-tion (Figure 1D, left), with similar branch numbers as the con-trols but a significantly larger cluster area. RTEF-1 dominant negative HMEC-1, lacking the activation domain of RTEF-1 and competing with endogenous RTEF-1, appeared to form disconnected structures (Figure 1D, left). Quantitative analysis of the branch numbers and cluster areas confirmed the visual observation (Figure 1D, right). Increased cell–cell connections in the Matrigel assay were not observed in our previous study using the same vector and techniques for other gene transfec-tions,11 which confirms this specific phenotype of RTEF-1.
Examination of these cells on Matrigel by a series of photos (Figure 1E) taken every 10 minutes within the same field revealed that RTEF-1 o/e HMEC-1 appeared to accelerate endothelial cell connections to form a continuous monolayer with a large cell aggregation area. Similar results were also obtained when HUVEC/RTEF-1 o/e cells were plated on Matrigel (Figure I in the online-only Data Supplement). In addition, to confirm that cell aggregation was induced by RTEF-1, shRNA targeting RTEF-1 was used to knock down RTEF-1 expression in HMEC-1/RTEF-1 o/e cells. Knockdown of RTEF-1 by shRNA significantly reduced the cell aggregation effect (Figure 1F).
RTEF-1 Increases Gap Junction Proteins
To understand the molecular mechanisms underlying the observed RTEF-1–induced endothelial aggregation, we per-formed microarray analysis and compared gene expression patterns in RTEF-1–silenced endothelial cells with wild-type controls. A clear pattern emerged in which cell–cell junction proteins were regulated by RTEF-1. These findings were vali-dated by quantitative real-time polymerase chain reaction to analyze endothelial junction molecules including tight junc-tions, gap juncjunc-tions, and adhesion molecule expression in HMEC-1/RTEF-1 o/e (Figure 2A). A consequence of RTEF-1 overexpression was a significant increase in transcripts coding
Figure 1. Related transcription enhancer factor-1 (RTEF-1)–enhanced endothelial connections in Matrigel. A, Quantitative real-time poly-merase chain reaction (qPCR) (left) and Western blotting (right) of human microvascular endothelial cells (HMEC)-1 cells showing knock-down of RTEF-1 expression by 2 small interfering RNAs (siRNAs) in both mRNA and protein levels. qPCR data were from 3 independent experiments performed in triplicate and normalized to a housekeeping gene 36b4 (n=3) on plastic plates. Vinculin was used as a loading control for Western blotting. B, Images (left) and quantification (right) of vascular structure formation in Matrigel by wild-type HMEC-1 and RTEF-1 siRNA or control siRNA-treated HMEC-1 at 6 hours. The arrows point to cell–cell connections. Results are based on 3 indepen-dent experiments. *P<0.05. C, Schematic structure of RTEF-1 overexpression (o/e) and dominant negative (DN) cDNA constructs (upper). Western blotting images showing transgene expression in HMEC-1 cells on plastic plates (middle and lower). D, Images (left) from a Matrigel assay of wild-type HMEC-1, HMEC-1/forced overexpression of RTEF-1 (HMEC-1/RTEF-1 o/e) and HMEC-1/RTEF-1 dominant negative (DN) at 6 hours. Quantifications of branch numbers (right upper) showed a significant difference between HMEC-1/RTEF-1 DN and wild-type HMEC-1 but not HMEC-1/RTEF-1 o/e from wild-type HMEC-1, whereas larger clustered areas (right lower) were formed in HMEC-1/RTEF-1 o/e but not in HMEC-1/RTEF-1 DN cells. Results are based on 3 independent experiments. Endothelial network struc-tures were analyzed with Image J software, and results were expressed as mean±SEM. *P<0.05. E, Continuous observation of HMEC-1/ RTEF-1 o/e and wild-type HMEC-1 for 5 hours of the same 20× fields on Matrigel showing cell aggregation. F, Knockdown of RTEF-1 by 2 short hairpin RNA (shRNAs) in HMEC-1/RTEF-1 o/e cells was analyzed by Western blotting analysis (upper). Endothelial cell aggregation on Matrigel (lower left) was decreased in RTEF-1 shRNA-treated cells, and quantification (lower right). **P<0.01.
for a group of gap junction proteins (Cx37, Cx40, and Cx43), accompanied by increased protein levels. The other junction molecules were not affected by RTEF-1 expression (Figure 2A). These increased gap junction proteins were further detected in HMEC-1/RTEF-1 o/e in Matrigel, specifically the significant upregulation of Cx43 (Figure 2B). In addi-tion, within endothelial cells in which RTEF-1 was knocked down by siRNA, downregulation of gap junction proteins was observed, and this was confirmed by quantitative real-time polymerase chain reaction and Western blotting analysis (Figure 2C). These results indicate that gap junction pro-teins, particularly Cx43, were upregulated by RTEF-1, which suggests that these proteins might play an important role in RTFE-1–driven endothelial cell connections and aggregation. RTEF-1 Stimulates Functional Connexins to Enhance Endothelial Cell Connections and Aggregation
Among the connexins, Cx43 has been considered a major gap junction protein, and Cx43 was predominantly induced by
RTEF-1 both in normal culture and in Matrigel. To determine whether RTEF-1 stimulated Cx43 at the transcriptional level, the activity of a luciferase construct under the control of the Cx43 promoter was measured. The Cx43-specific luciferase expression increased 2.25-fold in HMEC-1/RTEF-1 o/e compared with HMEC-1/control (Figure 3A). The proximal promoters for the mouse, human, and rat Cx43 genes have been mapped in several Cx43-expressing cell types to an evolutionary conserved region of Sp-binding sites and activator protein 1–binding elements.19 These Sp1 and activator protein 1–binding sites were shown to contribute to promoter activity and to bind the transcription factors Sp1/Sp3 or activator protein 1 and activate the rat Cx43 promoter. Previously, we have shown that RTEF-1 regulates its target genes through the muscle-CAT element and can also transcriptionally regulate VEGF through its Sp1 site.8 To further investigate these Sp1 sites as key elements in RTEF-1–regulating Cx43 transcription, a chromatin immunoprecipitation assay was performed (Figure 3B). Primers were designed to flank an Sp1 sequence of Cx43, and the chromatin immunoprecipitation
Figure 2. Related transcription enhancer factor-1 (RTEF-1) increases gap junction proteins. A, Quantitative real-time polymerase chain reaction (qPCR) (left), Western blotting (middle), and Western blot quantification (right) indicate the expression of cell adhesion molecules, tight junction molecules, and gap junction molecules in human microvascular endothelial cells-1/forced overexpression of RTEF-1 (HMEC-1/ RTEF-1 o/e) cells compared with control HMEC-1 cells grown on plastic plates. qPCR data were from 3 independent experiments per-formed in triplicate and normalized to housekeeping gene 36b4 (n=3). To quantitate the Western blot images, Image J was used to measure pixel density before normalizing with the loading control (Vinculin), and compared with protein expression in the wild-type HMEC-1. Results are mean±SD based on 3 experiments. *P<0.05, **P<0.01. B, qPCR (left) indicates the expression of gap junction molecules in endothelial cells in Matrigel. Western blotting (right) analysis shows connexin (Cx) 37, Cx40, and Cx43 proteins in HMEC-1/RTEF-1 o/e in the Matrigel compared with control HMEC-1. mRNA and protein were extracted from cell lysates obtained from dissolved cell networks in Matrigel. Both the mRNA and protein levels of gap junctions are increased in RTEF-1 o/e HMEC-1 in Matrigel. *P<0.05. C, qPCR (left) and Western blotting (right) indicates the expression of gap junction molecules in HMEC-1 cells in which RTEF-1 was knocked down with small interfering RNA (siRNA) compared with control HMEC-1 cells grown on plastic plates. Both the mRNA and protein levels of gap junctions are decreased in RTEF-1 siRNA-transfected HMEC-1. Results are mean±SD based on 3 experiments. *P<0.05.
A
B C
Figure 3. Related transcription enhancer factor-1 (RTEF-1) stimulates functional connexins to enhance endothelial cell connections and aggregation. A, Luciferase assays indicate that connexin 43 (Cx43) promoter activity was increased in human microvascular endothelial cells-1/forced overexpression of RTEF-1 (HMEC-1/RTEF-1 o/e). Results are mean±SD. B, Chromatin immunoprecipitation assay dem-onstrating the RTEF-1 antibody immunoprecipitating with the specificity protein 1 region of the Cx43 promoter. C, Immunostaining with Cx43 antibody (red) and 4′,6-diamidino-2-phenylindole (DAPI, blue) demonstrate Cx43 in HMEC-1/RTEF-1 o/e and wild-type HMEC-1 grown on plastic plates. More Cx43-positive staining was found in the HMEC-1/RTEF-1 o/e. Scale bar=20 μm. D, Scrape loading image (left). Dye spread in images (dye intensity) was quantified using Image J software (right). *P<0.05 (Results of 3 independent experiments). E, Western blot (left) showing knockdown of Cx43 by 2 different sequences of Cx43 siRNA. Images (middle) and quantification (right) of disconnected endothelial networks formed by HMEC-1 with endogenous Cx43 knockdown by 2 different sequences of Cx43 siRNA. *P<0.05. F, HMEC-1/RTEF-1 o/e were treated with negative control siRNA or 2 different sequences of Cx43 siRNA for 48 hours and then plated in Matrigel for 6 hours. RTEF-1–induced cell aggregation areas (arrows) were diminished in Cx43 siRNA-treated HMEC-1/RTEF-1 o/e. Images are represented as 3 independent experiments (upper) and quantification is on the lower left. RTEF-1 and Cx43 mRNA levels were quantified by quantitative real-time polymerase chain reaction in wild-type and Cx43 siRNA treated HMEC-1 cells grown on plastic plates (bottom right). *P<0.05; **P<0.01.
A C E F D B
assay demonstrated that RTEF-1 specifically bound to the Cx43 promoter.
Immunofluorescence assay confirmed increased Cx43 expression in HMEC-1/RTEF-1 o/e cells (Figure 3C). This increased Cx43 expression was found to be localized on the endothelial cell membrane (Figure II in the online-only Data Supplement). Considering that endothelial cell–cell connec-tions, mediated primarily by gap junctions including con-nexins, are important in maintaining ionic and metabolic homeostasis in endothelial cells, a scrape-loading assay was performed to assess whether the changes in the connexins of HMEC-1/RTEF-1 o/e correlated with the potential cell–cell coupling function. Quantification of dye spread indicated that increased expression of connexins was associated with a signif-icant increase in gap junction coupling in HMEC-1/RTEF-1 o/e cells (Figure 3D). To test whether the RTEF-1–driven increase in Cx43 expression was sufficient to permit cellular connec-tions and aggregation, we knocked down Cx43 in HMEC-1 cells with Cx43 siRNA before the cells were plated in Matrigel (Figure 3E). Knockdown of Cx43 by siRNA in wild-type HMEC-1 resulted in a disconnected network structure (Figure 3E), which was very similar to the phenotype of endogenous RTEF-1 knockdown cells in Matrigel. In addition, recent studies involving knockdown of Cx43 by targeted siRNA in endothelial cells demonstrated no effect on proliferation, con-firming that the accumulation of Cx43 induced by RTEF-1 results in cell aggregation.20 Furthermore, cluster areas formed by RTEF-1–enhanced cell aggregation in Matrigel were signif-icantly reduced after Cx43 siRNA treatment and, interestingly, Cx43 siRNA also attenuated RTEF-1 mRNA levels (Figure 3F). We have previously determined that RTEF-1 upregulates VEGF-B expression, so we examined whether VEGF-B signal-ing also contributes to RTEF-1–induced aggregation. However, as shown in Figure III in the online-only Data Supplement, we blocked the VEGF receptor and did not see significant changes in RTEF-1–induced endothelial aggregation.
RTEF-1 Is Critical for Vascular Structure Formation In Vivo
To explore the effect of RTEF-1 in vascular structure formation in vivo, we generated mice with RTEF-1 driven by the tetra-cycline-mediated VE-Cadherin promoter (VE-Cad:RTEF-1), which specifically expressed RTEF-1 in endothelial cells (Figure IV in the online-only Data Supplement). In normoxic conditions, retinal capillaries of VE-Cad:RTEF-1 mice devel-oped normally. To better define the role of RTEF-1 during mice retina vascularization, an OIR model was used, in which the function of RTEF-1 can be detected in the induced path-ological neovascularization. In the OIR model, exposure of mouse pups to hyperoxia (75% oxygen) for 5 days from P7 to P12 resulted in a rapid obliteration of capillaries in a pattern that is most pronounced near the central optic disc, and vessel regrowth and neovascularization are triggered after returning to ambient air. In analyzing this experiment, we focused on neovascularization caused by hypoxia upon the mice return-ing to the ambient air after hyperoxgen-induced vessel regres-sion. We observed that the retinas of VE-Cad:RTEF-1 mice indicated excessive angiogenesis and a significantly reduced avascular area compared with wild-type littermates on P17
(Figure 4A) when the neovascularization reached its maxi-mum. Analysis indicated a 40% decrease in the avascular area in VE-Cad:RTEF-1 mouse retina (Figure 4A). In addi-tion to excessive neovascularizaaddi-tion, other morphological changes were observed in VE-Cad:RTEF-1 mice. A denser, more highly branched vascular plexus was observed in VE-Cad:RTEF-1 mice in comparison with littermate controls (Figure 4B). Analysis indicated a 28% increase in isolectin-positive staining in the VE-Cad:RTEF-1 mice.
To ensure that the morphology changes in the RTEF-1 mouse retina were because of the induction of connexins, we further evaluated RTEF-1 and Cx43 expression in retinas from VE-Cad:RTEF-1 mice. As noted in Figure 4C, Western blots for Cx43 of whole retinal lysates from VE-Cad:RTEF-1 showed an increase on P8 in retina as well as in adult hearts. Thus, RTEF-1 enhances angiogenic sprouting and promotes the formation of a denser primary capillary network with an upregulation of Cx43 in vivo.
Discussion
We previously reported that the transcription factor RTEF-1 promotes angiogenesis by upregulating VEGF to enhance endothelial cell proliferation.8 The present report demon-strates another important role of RTEF-1 in vascular struc-ture formation. Using complimentary approaches to knockout or force-overexpress RTEF-1 in various physiological and pathological settings, we provide several lines of evidence that RTEF-1 is one of the key transcriptional regulators that controls the endothelium to form a network structure. First, we found that RTEF-1 promoted endothelial network forma-tion through increasing cell–cell connecforma-tions and aggregaforma-tion; without endogenous RTEF-1, the endothelial network became scattered and disconnected. Second, the precise mechanism by which RTEF-1 regulates vascular structure formation was demonstrated through stimulating gap junction proteins; spe-cifically, we demonstrated that Cx43 plays an important role in RTEF-1–driven endothelial network formation. Third, we observed enhanced endothelial sprouting and a denser vas-cular structure during neovasvas-cularization in the OIR model in RTEF-1–overexpressing mice. These data indicate that RTEF-1 expression is important in the formation of cell–cell connections required to form a capillary network.
Capillary network formation is a complex process, and sev-eral genes can be coordinately involved. RTEF-1, as a tran-scription factor, is capable of regulating these genes to control the process, particularly by accelerating endothelial cell con-nections and aggregation. Endothelial cell concon-nections and aggregation can occur in angiogenesis during tissue repair or in disease settings such as cancer and diabetic retinopathy. Other transcription regulators, such as the basic helix-loop-helix transcription factor (hairy and enhancer of split related-1), are necessary for the induction of tubular network formation via downregulation of vascular endothelial growth factor receptor 2, which suggests that hairy and enhancer of split related-1 regulates the transition from the proliferating, migrating phe-notype to the network-forming and vessel maturation pheno-type.21 Another transcription factor, Erg, has been reported to regulate angiogenesis and endothelial apoptosis through the junctional adhesion molecule VE-cadherin.22 The transcription
factor T-cell acute lymphocytic leukemia protein 1/stem cell leukemia regulates endothelial tube formation also by target-ing VE-cadherin.23 We suggest that RTEF-1 is involved in the expression of gap junction proteins and contributes to the pre-cise regulation of connexins at the level of cell–cell connec-tions during vascular development and angiogenesis.
Endothelial cells must form cell–cell contacts to form capillary-like networks, and cell–cell adhesion is mediated by several proteins including tight junctions, cell adhesion molecules, and gap junctions. We found that RTEF-1 specifically increased gap junction proteins at a transcriptional level via induction of Cx43 promoter activity, yet did not affect tight junction proteins or cadherin levels. Direct intercellular communication through gap junctions is critical in the control and coordination of vascular function. In the cardiovascular system, gap junctions are composed of 4 connexin proteins: Cx37, Cx40, Cx43, and Cx45. Vascular connexins work together; no single gap junction protein in the vasculature is
redundant during the development of the cardiovascular system or for integration of smooth muscle and endothelial cells to synchronize cell functions along the length of the vessel wall.24,25 However, the mechanism of transcriptional control of these connexins in endothelium remains unclear. When Cx43 was knocked down in RTEF-1 o/e cells, RTEF-1–induced cell aggregation was significantly decreased, indicating that RTEF-1–induced cell aggregation was attributable to an increase in Cx43. RTEF-1 not only induced higher expression levels of connexins but also induced the formation of functional gap junctions in which connexin-based channels have emerged as an important signaling pathway.25 Moreover, a similar change in this endothelial cell behavior occurs in glioma cells, where Cx43 induced morphological transformation of glioma cells into an epithelial phenotype, which was associated with an increase in cell aggregation.26 In transgenic mice generated with cardiac-specific overexpression of RTEF-1, conduction defects were detected and correlated with the dephosphorylation of
Figure 4. Related transcription enhancer factor-1 (RTEF-1) is critical for vascular structure formation in vivo. A, Left: Griffonia simplicifolia lectin staining of oxygen-induced retinopathy (OIR) mouse retinas at postnatal day 17 (P17) from Control (A-a) and RTEF-1 mice driven by the tetracycline-mediated VE-Cadherin promoter (VE-Cad:RTEF-1) (A-b) mice. The retina from VE-Cad:RTEF-1 mice demonstrated excessive angiogenesis and reduced avascular area when compared with the control (A-a and A-b; magnification, 4×). A higher magni-fication of the avascular area is also shown (A-c and A-d; magnimagni-fication, 5×). Right: Analysis of avascular areas. Results of the analysis indicated a 40% decrease in the avascular area within the retina of RTEF-1 o/e mice compared with controls. n=4, *P<0.05. B, Left: Neovascular structure within the retina of control and RTEF-1 o/e mice on P17 after OIR. Images show capillary network induced by OIR in control retinas (B-a) and RTEF-1 o/e retinas (B-b). Images are shown at a magnification of 5×. B-c and B-d are higher magnification of the white rectangle areas (neovascularization area) in B-a and B-b to clearly show capillary network density and branching (magnification, 20×). Right: Analysis of the isolectin-positive area of B-c and B-d. n=4, *P<0.05. C, Western blot of the whole neonatal retina and adult hearts of VE-Cad:RTEF-1 mice showed the protein levels of connexin 43 were increased in VE-Cad:RTEF-1 mice compared with their litter mate controls.
A
B C
Cx40 and Cx43.27 However, in endothelium, our data indicated that both expression and activation of gap junction proteins were regulated by RTEF-1.
Diabetes mellitus–induced Cx43 downregulation promotes retinal vascular lesions characteristic of diabetic retinopathy and has been found in streptozotocin-induced diabetic mice and in Cx43 heterozygous knockout (Cx43+/−) mice.28 This indicates that reduced Cx43 expression plays a critical role in the development of acellular capillaries. We observed that the retinas of VE-Cad:RTEF-1 mice demonstrated excessive angiogenesis and a significantly reduced avascular area com-pared with wild-type littermate controls in an OIR model. It will be interesting to characterize the metabolic effect of delayed recovery from an ischemic area and abnormal vas-cular structure in the endothelial-deficient RTEF-1 mice to understand how the downregulation of connexins by RTEF-1 is related to retinal vascular lesions of diabetic retinopathy.
In conclusion, we have identified a new transcriptional pathway important in regulating vascular structure formation, which involves the transcription factor, RTEF-1, and the gap junction proteins, connexins. Developing the means for con-trolling endothelial connections is essential for finding new therapeutic targets in human vascular diseases. This finding may be relevant to pathological angiogenesis associated with various disease conditions and could be extended to investiga-tions of other mechanisms of angiogenesis.
Acknowledgments
We thank Dr Laura Benjamin (BIDMC, Harvard) for the VE-Cad:tTA transgenic mice, Dr Alexandre Stewart (University of Ottawa) for the PXJ40/RTEF-1 construct, and Drs Lye and Jiang (University of Texas) for Cx43 promoter construct.
Sources of Funding
This work was supported, in part, by National Institutes of Health grant HLR01082837 (to J.L.) and China Scholarship Council (to X.A. and Y.J.).
Disclosures
None.
References
1. Dejana E, Taddei A, Randi AM. Foxs and Ets in the transcriptional regulation of endothelial cell differentiation and angiogenesis. Biochim
Biophys Acta. 2007;1775:298–312.
2. Kuhnert F, Campagnolo L, Xiong JW, Lemons D, Fitch MJ, Zou Z, Kiosses WB, Gardner H, Stuhlmann H. Dosage-dependent require-ment for mouse Vezf1 in vascular system developrequire-ment. Dev Biol. 2005; 283:140–156.
3. Kappel A, Schlaeger TM, Flamme I, Orkin SH, Risau W, Breier G. Role of SCL/Tal-1, GATA, and ets transcription factor binding sites for the regulation of flk-1 expression during murine vascular development.
Blood. 2000;96:3078–3085.
4. Fischer A, Schumacher N, Maier M, Sendtner M, Gessler M. The Notch target genes Hey1 and Hey2 are required for embryonic vascular devel-opment. Genes Dev. 2004;18:901–911.
5. Yagi R, Kohn MJ, Karavanova I, Kaneko KJ, Vullhorst D, DePamphilis ML, Buonanno A. Transcription factor TEAD4 specifies the trophecto-derm lineage at the beginning of mammalian development. Development. 2007;134:3827–3836.
6. Stewart AF, Larkin SB, Farrance IK, Mar JH, Hall DE, Ordahl CP. Muscle-enriched TEF-1 isoforms bind M-CAT elements from muscle-specific promoters and differentially activate transcription. J Biol Chem. 1994;269:3147–3150.
7. Larkin SB, Farrance IK, Ordahl CP. Flanking sequences modulate the cell specificity of M-CAT elements. Mol Cell Biol. 1996;16:3742–3755.
8. Shie JL, Wu G, Wu J, Liu FF, Laham RJ, Oettgen P, Li J. RTEF-1, a novel transcriptional stimulator of vascular endothelial growth factor in hypoxic endothelial cells. J Biol Chem. 2004;279:25010–25016. 9. Connor KM, Krah NM, Dennison RJ, Aderman CM, Chen J, Guerin
KI, Sapieha P, Stahl A, Willett KL, Smith LE. Quantification of oxy-gen-induced retinopathy in the mouse: a model of vessel loss, vessel regrowth and pathological angiogenesis. Nat Protoc. 2009;4:1565–1573. 10. Ades EW, Candal FJ, Swerlick RA, George VG, Summers S, Bosse DC,
Lawley TJ. HMEC-1: establishment of an immortalized human micro-vascular endothelial cell line. J Invest Dermatol. 1992;99:683–690. 11. An X, Jin Y, Guo H, Foo SY, Cully BL, Wu J, Zeng H, Rosenzweig A,
Li J. Response gene to complement 32, a novel hypoxia-regulated angio-genic inhibitor. Circulation. 2009;120:617–627.
12. Geimonen E, Jiang W, Ali M, Fishman GI, Garfield RE, Andersen J. Activation of protein kinase C in human uterine smooth muscle induces connexin-43 gene transcription through an AP-1 site in the promoter sequence. J Biol Chem. 1996;271:23667–23674.
13. Berdichevski A, Meiry G, Milman F, Reiter I, Sedan O, Eliyahu S, Duffy HS, Youdim MB, Binah O. TVP1022 protects neonatal rat ven-tricular myocytes against doxorubicin-induced functional derangements.
J Pharmacol Exp Ther. 2010;332:413–420.
14. De Pina-Benabou MH, Srinivas M, Spray DC, Scemes E. Calmodulin kinase pathway mediates the K+-induced increase in Gap junctional communication between mouse spinal cord astrocytes. J Neurosci. 2001;21:6635–6643.
15. Tirziu D, Chorianopoulos E, Moodie KL, Palac RT, Zhuang ZW, Tjwa M, Roncal C, Eriksson U, Fu Q, Elfenbein A, Hall AE, Carmeliet P, Moons L, Simons M. Myocardial hypertrophy in the absence of external stimuli is induced by angiogenesis in mice. J Clin Invest. 2007;117:3188–3197. 16. Meadows KN, Iyer S, Stevens MV, Wang D, Shechter S, Perruzzi C,
Camenisch TD, Benjamin LE. Akt promotes endocardial-mesenchyme transition. J Angiogenes Res. 2009;1:2.
17. Smith LE, Wesolowski E, McLellan A, Kostyk SK, D’Amato R, Sullivan R, D’Amore PA. Oxygen-induced retinopathy in the mouse. Invest
Ophthalmol Vis Sci. 1994;35:101–111.
18. Lebrin F, Srun S, Raymond K, Martin S, van den Brink S, Freitas C, Bréant C, Mathivet T, Larrivée B, Thomas JL, Arthur HM, Westermann CJ, Disch F, Mager JJ, Snijder RJ, Eichmann A, Mummery CL. Thalidomide stimulates vessel maturation and reduces epistaxis in individuals with hereditary hemorrhagic telangiectasia. Nat Med. 2010;16:420–428. 19. Teunissen BE, Bierhuizen MF. Transcriptional control of myocardial
connexins. Cardiovasc Res. 2004;62:246–255.
20. Johnson TL, Nerem RM. Endothelial connexin 37, connexin 40, and connexin 43 respond uniquely to substrate and shear stress. Endothelium. 2007;14:215–226.
21. Henderson AM, Wang SJ, Taylor AC, Aitkenhead M, Hughes CC. The basic helix-loop-helix transcription factor HESR1 regulates endothelial cell tube formation. J Biol Chem. 2001;276:6169–6176.
22. Birdsey GM, Dryden NH, Amsellem V, Gebhardt F, Sahnan K, Haskard DO, Dejana E, Mason JC, Randi AM. Transcription factor Erg regulates angiogenesis and endothelial apoptosis through VE-cadherin. Blood. 2008;111:3498–3506.
23. Deleuze V, Chalhoub E, El-Hajj R, Dohet C, Le Clech M, Couraud PO, Huber P, Mathieu D. TAL-1/SCL and its partners E47 and LMO2 up-regulate VE-cadherin expression in endothelial cells. Mol Cell Biol. 2007;27:2687–2697.
24. Bény JL. Information Networks in the Arterial Wall. News Physiol Sci. 1999;14:68–73.
25. Figueroa XF, Duling BR. Gap junctions in the control of vascular func-tion. Antioxid Redox Signal. 2009;11:251–266.
26. Lin JH, Takano T, Cotrina ML, Arcuino G, Kang J, Liu S, Gao Q, Jiang L, Li F, Lichtenberg-Frate H, Haubrich S, Willecke K, Goldman SA, Nedergaard M. Connexin 43 enhances the adhesivity and mediates the invasion of malignant glioma cells. J Neurosci. 2002;22:4302–4311. 27. Chen HH, Baty CJ, Maeda T, Brooks S, Baker LC, Ueyama T, Gursoy
E, Saba S, Salama G, London B, Stewart AF. Transcription enhancer factor-1-related factor-transgenic mice develop cardiac conduction defects associated with altered connexin phosphorylation. Circulation. 2004;110:2980–2987.
28. Bobbie MW, Roy S, Trudeau K, Munger SJ, Simon AM, Roy S. Reduced connexin 43 expression and its effect on the development of vascular lesions in retinas of diabetic mice. Invest Ophthalmol Vis Sci. 2010;51:3758–3763.
