COUP-TFII regulates the functions of Prox1 in lymphatic endothelial cells through direct interaction

Authors

  • Tomoko Yamazaki,

    1. Department of Molecular Pathology, Graduate School of Medicine, and the Global Center of Excellence Program for “Integrative Life Science Based on the Study of Biosignaling Mechanisms”, University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan
    Search for more papers by this author
  • Yasuhiro Yoshimatsu,

    1. Department of Molecular Pathology, Graduate School of Medicine, and the Global Center of Excellence Program for “Integrative Life Science Based on the Study of Biosignaling Mechanisms”, University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan
    Search for more papers by this author
  • Yasuyuki Morishita,

    1. Department of Molecular Pathology, Graduate School of Medicine, and the Global Center of Excellence Program for “Integrative Life Science Based on the Study of Biosignaling Mechanisms”, University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan
    Search for more papers by this author
  • Kohei Miyazono,

    Corresponding authorSearch for more papers by this author
  • Tetsuro Watabe

    1. Department of Molecular Pathology, Graduate School of Medicine, and the Global Center of Excellence Program for “Integrative Life Science Based on the Study of Biosignaling Mechanisms”, University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan
    Search for more papers by this author

  • Communicated by: Masayuki Yamamoto (Tohoku University)

* Correspondence: miyazono@m.u-tokyo.ac.jp

Abstract

During embryonic lymphatic development, Prox1 homeobox transcription factor is expressed in a subset of venous blood vascular endothelial cells (BECs) in which COUP-TFII orphan nuclear receptor is highly expressed. Prox1 induces differentiation of BECs into lymphatic endothelial cells (LECs) by inducing the expression of various LEC markers including vascular endothelial growth factor receptor 3 (VEGFR3). However, the molecular mechanisms of how transcriptional activities of Prox1 are regulated are largely unknown. In the present study, we show that COUP-TFII plays important roles in the regulation of the function of Prox1. In BECs and LECs, Prox1 promotes the proliferation and migration toward VEGF-C by inducing the expression of cyclin E1 and VEGFR3, respectively. Gain-of-function studies showed that COUP-TFII negatively regulates the effects of Prox1 in BECs and LECs whereas loss-of-function studies showed that COUP-TFII negatively and positively regulates Prox1 in BECs and LECs, respectively. We also show that endogenous Prox1 and COUP-TFII physically interact in LECs and that both Prox1 and COUP-TFII bind to the endogenous cyclin E1 promoter. These results suggest that COUP-TFII physically and functionally interact during differentiation and maintenance of lymphatic vessels.

Introduction

Lymphatic vascular systems play critical roles in the maintenance of tissue fluid homeostasis and the mediation of the afferent immune response. Defects in the lymphatic systems result in lymphedema. In pathological situations, they serve as routes of the metastatic spread of malignant tumors to regional lymph nodes. Because of such clinical relevance, understanding of the molecular mechanisms that govern lymphangiogenesis is crucial (Karpanen & Alitalo 2008).

Numerous groups have shown that activation of signaling pathways via vascular endothelial growth factor receptor 3 (VEGFR3) by VEGF-C/D plays central roles in the formation of lymphatic systems. Genetic ablation of Vegf-c gene leads to the lack of lymphatic formation (Karkkainen et al. 2004). Additionally, expression of VEGF-C under skin-specific promoter induces hyperplasia of cutaneous lymphatic vessels (Jeltsch et al. 1997; Veikkola et al. 2001). Furthermore, inhibition of VEGFR3 signals via VEGFR3-Fc trap leads to diminishment of lymphatic vessels (He et al. 2002).

However, lymphangiogenesis is not regulated only by VEGFR3 signaling pathways. Recent reports have shown that integrin α9/β1 complexes serve as receptors for VEGF-C/D to regulate cell migration (Vlahakis et al. 2005). Furthermore, receptor tyrosine kinases including Tie2 (Morisada et al. 2005), fibroblast growth factor receptor 3 (FGFR3: Shin et al. 2006), platelet-derived growth factor receptor β (PDGFRβ: Cao et al. 2004) and hepatocyte growth factor receptor (HGFR: Kajiya et al. 2005) have been implicated in lymphangiogenesis. Therefore, if there are transcription factors that regulate these multiple lymphangiogenic signals, such master regulators can be ideal candidates as targets of anti-lymphangiogenesis therapies.

During embryonic lymphatic development, a homeobox transcription factor Prox1 has been shown to play important roles in the differentiation of venous endothelial cells into lymphatic endothelial cells (LECs; Oliver 2004). At 9.5 dpc of mouse development, Prox1 starts to become expressed specifically in a subpopulation of endothelial cells located on one side of the anterior cardinal vein. At this stage, venous endothelial cells express CD34, a blood vascular endothelial cell (BEC) marker, and low level of VEGFR3, whose expression becomes restricted to LEC at later stages. Upon Prox1 expression, expression of BEC markers decreases while expression of LEC markers, such as podoplanin and VEGFR3, increases. These Prox1 expressing cells start sprouting from veins and migrate towards mesenchymal cells expressing VEGF-C. Importantly, in Prox1 deficient mice, the migration of LECs is arrested, leading to complete lack of lymphatic systems (Wigle & Oliver 1999; Wigle et al. 2002).

Being a transcription factor, Prox1 regulates the expression of various target genes. When Prox1 was adenovirally transduced into human dermal microvascular endothelial cells (HDMECs), expression of LEC-specific genes was up-regulated (Petrova et al. 2002). Although Prox1-mediated induction of LEC-specific genes was not observed in non-BECs, Prox1 was capable of inducing the expression of cyclin E1 and E2 in various cell types. These results suggest that Prox1 may induce cell proliferation and differentiation of BECs into the LECs.

We recently examined the effects of Prox1 on the migration of two types of endothelial cells, mouse embryonic stem cell-derived endothelial cells and human umbilical vein endothelial cells (HUVECs) (Mishima et al. 2007). Prox1 induces the expression of VEGFR3 and integrin α9, which results in the endothelial migration towards VEGF-C. Furthermore, when Prox1 expression was knocked-down in human dermal LECs (HDLECs), expression of VEGFR3 and integrin α9 was attenuated with decrease in the migration towards VEGF-C. These results suggest that Prox1 serves as a master regulator in the differentiation and maintenance of LECs.

However, the molecular mechanisms of how Prox1 regulates the transcription of its target genes have been poorly understood. Shin and colleagues showed that Prox1 directly binds to the FGFR3 promoter to induce its expression in endothelial cells (Shin et al. 2006). Prox1 has also been shown to bind to the βB1-crystallin promoter to regulate its expression in lens epithelium (Cui et al. 2004). While various transcription factors are involved in the regulation of βB1-crystallin expression, the roles of Prox1 binding proteins in the Prox1-mediated transcriptional regulation have not yet been elucidated.

Qin et al. showed that Prox1 binds liver receptor homologue-1 (LRH-1/NR5A2), a member of fushi tarazu factor 1 subfamily of orphan nuclear receptors, which positively regulates the expression of cholesterol 7-α-hydroxylase (cyp7a1) in liver. Prox1 negatively regulates the transcriptional activities of LRH-1 by sequestrating LRH-1 proteins from cyp7a1 promoter (Qin et al. 2004). The suppression of the transcriptional activities of LRH-1 by Prox1 does not require the DNA binding domain of Prox1. These results suggest that other nuclear receptor family members may also physically and functionally interact with Prox1.

Chicken ovalbumin upstream promoter transcription factors (COUP-TFs) are orphan members of the steroid/thyroid hormone receptor superfamily. Two genes termed COUP-TFI (also known as EAR3/NR2F1) and COUP-TFII (also known as ARP-1/NR2F2) are closely related members and are expressed in various organs. COUP-TFs play important roles in the regulation of organogenesis, neurogenesis, and cellular differentiation during embryonic development. In blood vessels, COUP-TFII is specifically expressed in venous but not in arterial endothelium (You et al. 2005). Targeted disruption of COUP-TFII in endothelial cells results in the acquisition of arterial characteristics in mutant veins, suggesting that COUP-TFII has a critical role in maintaining vein identity. As lymphatic vessels are originated from veins, ablation of COUP-TFII in endothelial cells causes the decrease in Prox1-expressing cells (Srinivasan et al. 2007). However, the roles of COUP-TFII in LECs have not yet been elucidated.

In the present study, we found that COUP-TFII is expressed in LECs. By gain- and loss-of-function analyses, we showed that COUP-TFII suppresses the transcriptional activities of Prox1 to induce VEGFR3 and cyclin E1 in HUVECs, which leads to the inhibition of Prox1-mediated induction of endothelial cell proliferation and migration towards VEGF-C. Interestingly, both gain- and loss-of-function of COUP-TFII in HDLECs suppressed the expression of VEGFR3 and cyclin E1, suggesting that endogenous level of COUP-TFII is required to maintain the characteristics of LECs. Furthermore, we showed that COUP-TFII physically interacts with Prox1 and that both COUP-TFII and Prox1 bind to the cyclin E1 promoter. These results suggest that COUP-TFII regulates the transcriptional activities of Prox1 in LECs via physical interaction.

Results

COUP-TFII is expressed in human LECs

We first studied the expression of COUP-TFII in BECs and LECs using HUVECs and HDLECs by Western blot analysis (Fig. 1A). While COUP-TFII protein was detected in both HUVECs and HDLECs, its expression level was lower in HDLECs, in which Prox1 was expressed. We also observed that COUP-TFII protein was localized to the nuclei of HUVECs (Fig. 1B) and co-localized with Prox1 in HDLECs (Fig. 1C). The specificity of anti-COUP-TFII antibody used was confirmed using the HUVECs whose expression of COUP-TFII was knocked down by siRNA (Fig. S1 in Supporting Information).

Figure 1.

COUP-TFII is expressed in BECs and LECs. (A) Western blot analyses for COUP-TFII (top panel) and Prox1 expressions (middle panel) in HUVECs and in HDLECs. α-tubulin was used as a loading control (bottom panel). (B–C) Immunostaining of HUVECs (B) and HDLECs (C) was carried out for COUP-TFII (red) and Prox1 (green) with nuclear staining by TOTO3 (blue). COUP-TFII and Prox1 were co-localized to nuclei in HDLECs (see Merge). Scale bars, 50 µm. (D) Immunohistochemistry was carried out for COUP-TFII (red) and LYVE-1 (green) with nuclear staining by TOTO-3 (blue) using transverse sections at the level of the heart of 11.5 dpc mouse embryo. Right small panels are magnified images of the boxed area of the left large panel. CV; cardinal vein. Scale bars, 100 µm (left panel) and 20 µm (right four panels).

Furthermore, we examined the COUP-TFII expression in embryos. At 11.5 dpc of mouse development, LECs that are positive for LYVE-1, a LEC marker, sprout out from cardinal veins (Fig. 1D) (Oliver 2004). We observed that these sprouting LECs express COUP-TFII (Fig. 1D). These results suggest that COUP-TFII is temporally and spatially co-localized with Prox1 in LECs.

COUP-TFII suppresses Prox1-induced cell proliferation through the regulation of cyclin E1 expression

While Prox1 was reported to induce cyclin E1 expression in HDMECs (Petrova et al. 2002), its effects on endothelial cell proliferation have not yet been examined. When Prox1 was expressed in HUVECs using adenovirus, it significantly increased cell number (Fig. 2A). In order to examine the effect of COUP-TFII on Prox1-mediated promotion of endothelial cell proliferation, we increased the level of COUP-TFII expression by adenovirus coding for COUP-TFII. While COUP-TFII expression itself did not affect cell proliferation, elevated cell proliferation by Prox1 was significantly repressed by COUP-TFII (Fig. 2A).

Figure 2.

COUP-TFII suppresses Prox1-induced cell proliferation through the regulation of cyclin E1. (A) Numbers of cells were counted 48 h after infection of HUVECs with adenovirus coding for Prox1 (Ad-Prox1) in combination with that for COUP-TFII (Ad-COUP-TFII). Each value represents the mean of triplicate determinations; Bars, SD. (B–C) Effects of gain- and loss-of-function of COUP-TFII on the expression of cyclin E1 in HUVECs. HUVECs were infected with Ad-Prox1 in combination with Ad-COUP-TFII (B) or siRNAs for COUP-TFII (C), followed by quantitative RT-PCR analysis for cyclin E1. Non-coding adenovirus (Null) and siRNA carrying scrambled sequences (NTC) were used as negative controls. Bars, SD.

In order to dissect the molecular mechanisms, we carried out quantitative RT-PCR analysis for cyclin E1 (Fig, 2B) and E2 (Fig. S2 in Supporting Information). In accordance with the result of cell proliferation, COUP-TFII significantly suppressed the cyclin E1 expression induced by Prox1 (Fig. 2B), which was also confirmed for cyclin E2 expression (Fig. S2A in Supporting Information).

We next examined whether endogenous COUP-TFII is necessary to suppress Prox1-mediated induction of cyclin E1 expression by knocking down endogenous COUP-TFII expression using siRNA (Fig. S1 in Supporting Information). As shown in Fig. 2C, the induction of cyclin E1 expression by Prox1 was significantly increased by the loss of COUP-TFII expression, which was also confirmed for cyclin E2 expression (Fig. S2B in Supporting Information). These findings suggest that COUP-TFII suppresses Prox1-induced cell proliferation by interfering with cyclin E expression.

COUP-TFII suppresses Prox1-mediated endothelial cell migration towards VEGF-C by regulating VEGFR3 expression

We recently showed that Prox1 induces endothelial cell migration towards VEGF-C through up-regulation of VEGFR3 expression (Mishima et al. 2007). To examine the effect of COUP-TFII on the Prox1-mediated promotion of endothelial chemotaxis towards VEGF-C, we carried out chamber migration assays using HUVECs. As shown in Fig. 3A, COUP-TFII significantly suppressed the chemotaxis towards VEGF-C enhanced by Prox1.

Figure 3.

COUP-TFII represses Prox1-induced endothelial cell migration towards VEGF-C via regulation of VEGFR3 expression. (A) Effect of COUP-TFII on the Prox1-induced endothelial cell migration towards VEGF-C. Migration of HUVECs infected with indicated adnoviruses was measured by Boyden chamber assay as described previously (Mishima et al. 2007). Relative migration towards VEGF-C is shown as a ratio of the number of migrated cells in the presence of VEGF-C against that in the absence of VEGF-C. Bars, SD. (B–C) Effect of gain-of-function of COUP-TFII on the Prox1-induced expression of VEGFR3. HUVECs were infected with indicated adenoviruses, followed by quantitative RT-PCR (B) and Western blotting (C: top panel) for VEGFR3. Western blotting was also carried out for FLAG-tagged Prox1 and COUP-TFII transduced by adenoviruses (middle panel). α-tubulin was used as a loading control (bottom panel). (D) Effect of loss-of-function of COUP-TFII on the Prox1-induced expression of VEGFR3. HUVECs were infected with adenovirus coding for Prox1 (Ad-Prox1) or non-conding adenovirus (Ad-Null) in combination with siRNA for COUP-TFII or negative control siRNA (NTC), followed by quantitative RT-PCR for VEGFR3. Bars, SD.

In consistent with the results of the chamber migration assay, COUP-TFII suppressed the VEGFR3 mRNA expression induced by Prox1 (Fig. 3B). This result was further confirmed at protein level (Fig. 3C). In addition, when COUP-TFII expression was knocked down in HUVECs, induction of VEGFR3 expression by Prox1 was enhanced (Fig. 3D). These results suggest that COUP-TFII suppresses Prox1-induced endothelial chemotaxis towards VEGF-C by regulating VEGFR3 expression.

Endogenous level of COUP-TFII expression in HDLECs is required to maintain the expression of VEGFR3

We next investigated the roles of COUP-TFII in the LECs in which endogenous Prox1 is expressed. When the level of COUP-TFII expression was elevated in HDLECs, both proliferation (Fig. 4A) and chemotaxis towards VEGF-C (Fig. 4C) were suppressed with concomitant decrease in the expression of cyclin E1 (Fig. 4B) and VEGFR3 (Fig. 4D). These findings were consistent with the results observed in HUVECs. However, when COUP-TFII was knocked down in HDLECs, their migration towards VEGF-C and VEGFR3 expression were attenuated (Fig. 4E,F), whereas neither cell number nor cyclin E1 expression changed (data not shown). These results suggest that COUP-TFII differentially functions between HDLECs and HUVECs.

Figure 4.

Endogenous level of COUP-TFII expression plays important roles in the maintenance of characteristics of HDLECs. (A–D) HDLECs infected with non-coding adenovirus (Ad-Null) or adenovirus coding for COUP-TFII (Ad-COUP-TFII) were subjected to cell proliferation assay (A), chamber migration assay using VEGF-C as an attractant (C) and quantitative RT-PCR analyses for cyclin E1 (B) and VEGFR3 (D). (E, F) HDLECs transfected with control siRNA (siNTC) or siRNA for COUP-TFII (siCOUP-TFII) were subjected to chamber migration assay using VEGF-C as an attractant (E) and quantitative RT-PCR analyses for VEGFR3 (F). In (C) and (E), relative migration towards VEGF-C is shown as a ratio of the number of migrated cells in the presence of VEGF-C against that in the absence of VEGF-C. (G, H) Expression of endogenous Prox1 expression was determined in the HDLECs infected with Ad-COUP-TFII (G) or transfected with siCOUP-TFII (H). *P < 0.01. Bars, SD.

We further examined whether COUP-TFII alters the endogenous Prox1 expression in LECs. As shown in Fig. 4G, endogenous Prox1 expression was significantly decreased when the level of COUP-TFII expression was elevated in HDLECs, while the loss of COUP-TFII expression did not alter the Prox1 expression (Fig. 4H). These results suggest that excessive level of COUP-TFII may interfere with the functions of Prox1 directly by inhibiting the transcriptional activities of Prox1 and indirectly by suppressing the endogenous Prox1 expression.

COUP-TFII interacts with Prox1 in HDLECs

Previous reports that LRH-1 and SF-1, members of nuclear receptor superfamily, bind Prox1 (Qin et al. 2004; Steffensen et al. 2004) prompted us to examine whether COUP-TFII and Prox1 interact. We carried out co-immunoprecipitation experiments with cell lysates prepared from the HUVECs infected with adenoviruses coding for Prox1 and COUP-TFII (Fig. 5A). When the lysates were subjected to immunoprecipitation, we detected COUP-TFII in the immunoprecipitates pulled down with anti-Prox1 antibody, which indicates that Prox1 is capable of interacting with COUP-TFII in HUVECs.

Figure 5.

COUP-TFII physically interacts with Prox1. (A) Co-immunoprecipitation of COUP-TFII with Prox1 in HUVECs. HUVECs infected with adenovituses coding for Prox1 (Ad-Prox1) and COUP-TFII (Ad-COUP-TFII) were subjected to immunoprecipitation with anti-Prox1 antibody, followed by immunoblotting with anti-COUP-TFII antibody (top panel). Expression of Prox1 and COUP-TFII was also confirmed. (B) Interaction of endogenous Prox1 and COUP-TFII in HDLECs. PLA was carried out to detect the proximal location of Prox1 and COUP-TFII (shown as red dots) as described in Experimental Procedures. All samples were counterstained with TOTO3 (blue) to visualize nuclei. (a–c) HUVECs infected with adenoviruses coding for Prox1 and COUP-TFII (a), native HDLECs (b) and native HUVECs (c) were subjected to PLA after treating with antibodies for Prox1 and COUP-TFII. Note that specific interaction between Prox1 and COUP-TFII is detected in the nuclei only when Prox1 and COUP-TFII are present. (d) Native HDLECs were subjected to PLA without treating with antibodies for Prox1 and COUP-TFII. Scale bars, 10 µm.

Next, we examined whether endogenous Prox1 and COUP-TFII interact in HDLECs using the Duolink in situ proximity ligation assay (PLA). This method enables us to monitor subcellular localization of endogenous protein–protein interactions at single molecule resolution (Söderberg et al. 2006, 2008). In the HUVECs infected with adenoviruses coding for Prox1 and COUP-TFII, we detected a number of strong fluorescence signals in the presence of specific antibodies, which indicates the interaction between Prox1 and COUP-TFII in the nuclei (Fig. 5Ba), whereas no signals were detected in native HUVECs in the presence of specific antibodies (Fig. 5Bc) or in HDLECs in the absence of specific antibodies (Fig. 5Bd). In the native HDLECs, we could detect definite fluorescence signals restricted to the nuclei in the presence of specific antibodies (Fig. 5Bb), suggesting that endogenous Prox1 and COUP-TFII interact in the nuclei of HDLECs. In order to examine the specificity of the signals, we knocked down COUP-TFII expression by siRNA in HDLECs and carried out PLA (Fig. S3 in Supporting Information). The fluorescence signals seen in the HDLECs transfected with control siRNA were significantly decreased by knocking down COUP-TFII expression. These results allowed us to conclude that endogenous COUP-TFII interacts with Prox1 in the nuclei of HDLECs.

COUP-TFII and Prox1 bind to the cyclin E1 promoter

Petrova et al. showed that Prox1 activates cyclin E1 promoter whereas Prox1 DNA binding mutant does not (Petrova et al. 2002), suggesting that Prox1 may regulate the transcription of cyclin E1 via direct binding to the cyclin E1 promoter. Because COUP-TFII suppresses Prox1-induced cyclin E1 expression and physically interacts with Prox1, we examined whether Prox1 and COUP-TFII bind to the endogenous cyclin E1 promoter in intact chromatin.

Cross-linked chromatin samples prepared from HUVECs infected with adenoviruses coding for Prox1 and COUP-TFII were subjected to chromatin immunoprecipitation (ChIP) assays (Fig. 6). The cyclin E1 promoter region containing putative binding consensus sequences for Prox1 and COUP-TFII was pulled down with antibodies for Prox1 and COUP-TFII, suggesting that both Prox1 and COUP-TFII bind to the cyclin E1 promoter.

Figure 6.

COUP-TFII and Prox1 directly bind to the cyclin E1 promoter. HUVECs infected with adenoviruses coding for Prox1 and COUP-TFII were subjected to ChIP assay. PCR was carried out to detect the cyclin E1 promoter containing putative binding sequences for Prox1 and COUP-TFII. Prox1 Ab and COUP Ab lanes show amplification of target sequences within the immunoprecipitates (IP) using antibodies for Prox1 and COUP-TFII, respectively. Control IgG lanes show PCR amplification of samples precipitated with corresponding control IgG antibodies. Input lanes show amplification of total input DNA (+) or no DNA (–).

Discussion

In the present study, we show that COUP-TFII regulates the transcriptional activities of Prox1. In BECs, both gain- and loss-of-function studies showed that COUP-TFII negatively regulates Prox1 to induce the expression of cyclin E1 and VEGFR3. COUP-TFII has been reported to negatively regulate the transcription via binding to the promoter and recruitment of co-repressor complexes containing N-CoR, SMRT and histone deacetylase (HDAC) (Park et al. 2003). The present findings that COUP-TFII and Prox1 physically and functionally interact may suggest that positive transcriptional regulation by Prox1 is repressed by the co-repressor complexes that are recruited to the promoter by COUP-TFII.

We also found that endogenous level of COUP-TFII in LECs is required to maintain the expression of VEGFR3, which is not consistent with the results observed in HUVECs. This difference may be caused by the cell type specific contexts of expression of other transcription factors. Additionally, elevation of the level of COUP-TFII expression in LECs also decreased the expression of VEGFR3. We also found that both gain- and loss-of-function of COUP-TFII decreased the expression of LEC markers including integrin α9 and podoplanin and BEC markers including VE-cadherin and VEGFR2 (Fig. S4). These results suggest that a certain range of COUP-TFII expression is required to maintain the expression of a group of LEC and BEC markers. Similar phenomenon is observed in the relationship between the expression of Oct3/4 and maintenance of pluripotency of mouse embryonic stem cells (Niwa et al. 2000).

While endothelial markers examined so far appear to require endogenous level of COUP-TFII in HDLECs, we found that not all of endothelial markers are regulated by COUP-TFII in a similar manner. Ablation of COUP-TFII gene in endothelial cells allowed the ectopic expression of ephrin B2 and Neuropilin 1 (NRP1), both of which are arterial endothelial cell markers, in veins (You et al. 2005). In accordance with the previous result, COUP-TFII expression decreased ephrin B2 expression while loss-of-COUP-TFII expression increased it in HDLECs (Fig. S4E in Supporting Information). However, to our surprise, both gain- and loss-of-function studies showed that COUP-TFII positively regulates NRP1 expression in HDLECs (Fig. S4F in Supporting Information). Molecular mechanisms of how COUP-TFII differentially regulates the transcription of various target genes between BECs and LECs remain to be investigated in the future.

During embryonic lymphatic development, Prox1 expressing BECs in veins differentiate into LECs and sprout out to form primary lymphatic plexus (Oliver 2004). While we showed that COUP-TFII is expressed in HDLECs at lower level than in HUVECs, it remains to be elucidated how COUP-TFII expression is regulated during lymphatic development in embryos. Loss-of-function studies show that COUP-TFII negatively and positively regulates the VEGFR3 expression in BECs and LECs, respectively. These results may suggest that COUP-TFII maintains the identity of venous endothelial cells by inhibiting Prox1-mediated VEGFR3 expression in BECs, but aids in maintaining VEGFR3 expression in LECs. This differential transcriptional regulation by COUP-TFII may play important roles in segregating lymphatics from veins during formation of primary lymphatic plexus.

As Prox1 plays critical roles in the formation and maintenance of lymphatic vessels (Mishima et al. 2007), regulation of transcriptional activities of Prox1 will aid in manupilating lymphangiogenesis in pathological situations. While ligands for COUP-TFII receptor have not yet been identified, better understanding of the regulation of the COUP-TFII may be useful for developing therapeutic strategies to treat lymphedema and tumor lymphangiogenesis in the future.

Experimental procedures

Cell culture and adenovirus production

HUVECs and HDLECs were purchased from Sanko Junyaku and TaKaRa, and cultured in endothelial basal medium (EBM) containing 2% and 5% fetal bovine serum (FBS), respectively, supplemented with endothelial cell growth supplement (TaKaRa). Recombinant adenoviruses coding for mouse Prox1 and mouse COUP-TFII were generated and used as described (Shirakihara et al. 2007).

RNA interference and oligonucleotides

siRNAs were introduced into cells using HiperFect reagent (QIAGEN) according to the manufacturer's instructions. The COUP-TFII siRNA (SI00128814) and the negative control siRNA were obtained from QIAGEN.

Immunohistochemistry and Western blot analysis

Immunostaining was carried out with anti-Prox1 (1 : 100 dilution; Abcam), anti-COUP-TFII (1 : 100 dilution; Perseus Proteomics) and anti-LYVE-1 (1 : 200 dilution; Abcam) antibodies, followed by counterstaining with TOTO3 (Invitrogen–Molecular Probes). Stained specimens were examined using a LSM 510 META confocal microscope (Carl Zeiss). All images were imported into Adobe Photoshop as JPEGs or TIFFs for contrast manipulation and figure assembly. Antibodies to FLAG and α-tubulin for Western blot analysis were obtained from SIGMA. Antibody to human VEGFR3 for Western blot analysis was obtained from Santa Cruz. Western blot analysis was carried out as described (Watabe et al. 2003).

Proximity ligation assay (PLA)

The Duolink in situ PLA kits were purchased from Olink (http://www.olink.com/). Fixation of the cells, blocking of non-specific binding of antibody, and immunostaining using anti-Prox1 (Abcam) and anti-COUP-TFII (Perseus Proteomics) were carried out as described above. Subsequently, a pair of secondary antibodies conjugated with oligonucleotides (PLA probes) was used according to the manufacturer's protocol to generate fluorescence signals only when the two PLA probes were in close proximity (40 nm). The fluorescence signal from each detected pair of PLA probes was visualized as a distinct individual dot (Söderberg et al. 2006, 2008). Nuclei counterstaining and analysis of the images were carried out as described earlier.

Isolation of RNA and quantitative RT-PCR

Total RNAs were extracted from HUVECs and HDLECs using the RNeasy Mini Kit (QIAGEN). First-strand cDNAs were synthesized by SuperScriptIII reverse transcriptase (Invitrogen) using random hexamer primers according to the manufacturer's instruction. Quantitative RT-PCR analyses were carried out using the ABI PRISM 7500 Fast Real-Time PCR System (Applied Biosystems) and Power SYBR Green PCR master mix (Applied Biosystems). All expression data were normalized to those for β-actin. In the cases when siRNAs and adenoviruses were simultaneously used, expression data are presented as a ratio of the expression level in the samples infected with adenovirus coding for Prox1 against those with non-coding adenovirus. The primer sequences are available online as indicated in Table S1 in Supporting Information.

Chamber migration assay

Migration assay was carried out as described previously (Mishima et al. 2007). As a chemoattractant, 100 ng/mL and 300 ng/mL of recombinant VEGF-C (Calbiochem) were used for HDLECs and HUVECs, respectively.

ChIP assays

HUVECs infected with adenoviruses were fixed by adding formaldehyde and harvested. In order to precipitate Prox1 and COUP-TFII, anti-Prox1 antibody (R&D) and anti-COUP-TFII (Perseus Proteomics) were used. PCR of the cyclin E1 promoter containing putative binding sites for Prox1 and COUP-TFII was done using immunoprecipitated chromatin with the following pair of oligonucleotide primers:

5′-ACCAGCCTGAGCAACATAGCA-3′ and 5′-CAGTGAG ACCCCCATTTCTACA-3′.

Acknowledgements

Authors thank Drs. M. Tsai and S. Tsai for COUP-TFII cDNA, Ms. M. Arase for technical assistance, Drs. T. Minami, T. Kodama, and members of Department of Molecular Pathology of the University of Tokyo for discussion. This research was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Ancillary