Vascular Endothelial Growth Factor Up-regulates Expression of Receptor Activator of NF-κB (RANK) in Endothelial Cells

Vascular endothelial growth factor (VEGF) is known as a key regulator of angiogenesis during endochondral bone formation. Recently, we demonstrated that TNF-related activation-induced cytokine (TRANCE or RANKL), which is essential for bone remodeling, also had an angiogenic activity. Here we report that VEGF up-regulates expression of receptor activator of NF-κ B (RANK) and increases angiogenic responses of endothelial cells to TRANCE. Treatment of human umbilical vein endothelial cells (HUVECs) with VEGF increased both RANK mRNA and surface protein expression. Although placenta growth factor specific to VEGF receptor-1 had no significant effect on RANK expression, inhibition of downstream signaling molecules of the VEGF receptor-2 (Flk-1/KDR) such as Src, phospholipase C, protein kinase C, and phosphatidylinositol 3′-kinase suppressed VEGF-stimulated RANK expression in HUVECs. Moreover, the MEK inhibitor PD98059 or expression of dominant negative MEK1 inhibited induction of RANK by VEGF but not the Ca2+ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid-acetoxymethyl ester (BAPTA-AM). VEGF potentiated TRANCE-induced ERK activation and tube formation via RANK up-regulation in HUVECs. Together, these results show that VEGF enhances RANK expression in endothelial cells through Flk-1/KDR-protein kinase C-ERK signaling pathway, suggesting that VEGF plays an important role in modulating the angiogenic action of TRANCE under physiological or pathological conditions.


Vascular endothelial growth factor (VEGF) is known as a key regulator of angiogenesis during endochondral bone formation. Recently, we demonstrated that TNFrelated activation-induced cytokine (TRANCE or RANKL), which is essential for bone remodeling, also had an angiogenic activity. Here we report that VEGF up-regulates expression of receptor activator of NF-B (RANK) and increases angiogenic responses of endothelial cells to TRANCE. Treatment of human umbilical vein endothelial cells (HUVECs) with VEGF increased
both RANK mRNA and surface protein expression. Although placenta growth factor specific to VEGF receptor-1 had no significant effect on RANK expression, inhibition of downstream signaling molecules of the VEGF receptor-2 (Flk-1/KDR) such as Src, phospholipase C, protein kinase C, and phosphatidylinositol 3-kinase suppressed VEGF-stimulated RANK expression in HUVECs. Moreover, the MEK inhibitor PD98059 or expression of dominant negative MEK1 inhibited induction of RANK by VEGF but not the Ca 2؉ chelator 1,2-bis(2aminophenoxy)ethane-N,N,N,N-tetraacetic acidacetoxymethyl ester (BAPTA-AM). VEGF potentiated TRANCE-induced ERK activation and tube formation via RANK up-regulation in HUVECs. Together, these results show that VEGF enhances RANK expression in endothelial cells through Flk-1/KDR-protein kinase C-ERK signaling pathway, suggesting that VEGF plays an important role in modulating the angiogenic action of TRANCE under physiological or pathological conditions.
Angiogenesis, the development of new blood vessels from pre-existing endothelium, is a critical process in many physiological and pathological conditions including embryonic development, organ regeneration, chronic inflammation, and solid tumor growth (1,2). In particular, angiogenesis is a crucial first step in endochondral ossification, the process whereby cartilage is replaced by bone, which is essential for the development of the long bones (3). These angiogenic processes are controlled by a wide variety of positive or negative regulators, which are composed of growth factors, cytokines, lipid metabolites, and cryptic fragments of hemostatic proteins (4). Many studies have shown that various angiogenic and anti-angiogenic factors are secreted by chondrocytes presented in the growth plate region (5). Among these molecules, vascular endothelial growth factor (VEGF), 1 a potent mitogen for endothelial cells, has been best characterized for its role in the vascularization of bone tissues. VEGF is highly expressed in chondrocytes in the lower hypertrophic and mineralized regions of the cartilage in developing embryonic bones (6,7). The functional significance of VEGF for bone formation was determined in animal studies. Administration of a soluble VEGF receptor chimeric protein in juvenile mice almost completely suppressed blood vessel invasion at the growth plate and inhibited endochondral bone formation (8). Recently, studies using mice expressing exclusively the VEGF 120 isoform (VEGF 120/120 mice) have shown that loss of the VEGF 164 and VEGF 188 isoforms completely disturbed bone vascularization, and concomitantly resulted in a decreased trabecular bone volume and an abnormal growth plate morphology (9). Furthermore, in VEGF 120/120 bones differentiation of osteoblasts and hypertrophic chondrocytes was significantly decreased and the initial invasion of osteoblasts and endothelial cells into the bones was retarded (9). It is also reported that VEGF acts as a potent chemoattractant for cultured osteoblasts and stimulates the formation, survival, and 1 The abbreviations used are: VEGF, vascular endothelial growth factor; PlGF, placenta growth factor; TRANCE, tumor necrosis factorrelated activation-induced cytokine; HUVEC, human umbilical vein endothelial cell; RANK, receptor activator of NF-B; bFGF, basic fibroblast growth factor; EGF, epidermal growth factor; HGF, hepatocyte growth factor; IL, interleukin; KDR, Flk-1/kinase-insert domain-containing receptor; PI3K, phosphatidylinositol 3Ј-kinase; PLC, phospholipase C; PKC, protein kinase C; MEK, mitogen-activated protein/extracellular signal-regulated kinase kinase; ERK, extracellular signalregulated kinase; eNOS, endothelial nitric-oxide synthase; PP1, 4-amino-5-[4-methylphenyl]-7-[t-butyl]pyrazolo [3,4-d]pyrimidine; L-NAME, N G -nitro-L-arginine methyl ester; L-NAC, N-acetyl-L-cysteine; RT, reverse transcriptase; FBS, fetal bovine serum; HA, hemagglutinin; MAPK, mitogen-activated protein kinase; BAPTA-AM, 1,2-bis(2aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid tetrakis (acetoxymethyl ester); DN, dominant negative; FACS, fluorescence-activated cell sorting; TGF, transforming growth factor; TNF, tumor necrosis factor; PBS, phosphate-buffered saline; FITC, fluorescein isothiocyanate. resorption activity of osteoclasts in vitro (10 -12).
New blood vessel formation is crucial for establishing the conduit that allows a variety of cells essential for bone morphogenesis such as chondroclasts, osteoblasts, and osteoclasts to migrate into the growth plate. To develop and maintain normal bone tissues, osteoblastic matrix deposition and osteoclastic resorption must be closely coordinated. A perturbation of this balance can result in skeletal abnormalities characterized by decreased (osteoporosis) or increased (osteopetrosis) bone mass. Osteoblasts/stromal cells have been suggested for their role in the regulation of osteoclast maturation and activation (13). Recently, it was found that TNF-related activation-induced cytokine (TRANCE, also known as OPGL and RANKL) expressed on osteoblasts plays a key role in osteoclast differentiation from hematopoietic precursors and calcium metabolism (14,15). In a T cell-dependent model of rat adjuvant arthritis, TRANCE was also shown to be responsible for bone loss and cartilage destruction (16). These effects of TRANCE are exerted by its binding to the transmembrane receptor RANK (receptor activator of NF-B). RANK is shown to be detectable on mature dendritic cells, chondrocytes, osteoclast precursors, and mature osteoclasts (17,18), and induces activation of extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase, AKT, and NF-B in dendritic cells and osteoclasts (19 -21). In a recent study, we have demonstrated that TRANCE has a novel action of promoting angiogenesis in vivo, and the binding of TRANCE to RANK expressed on endothelial cells directly stimulates proliferation, migration, and tube formation of endothelial cells in vitro (22). Thus, these observations suggest that the TRANCE-RANK system may be important for the processes of osteoclastogenesis and vascularization of the growth plate during bone formation and remodeling.
Because VEGF is the angiogenic factor most closely associated with induction of the neovasculature in bone and also involved in osteoclastogenesis, it is suggested that VEGF may be functionally related with the biological action of TRANCE. In this study, we investigate the potential role of VEGF in the regulation of the TRANCE receptor RANK expression. We here show that VEGF increases RANK expression in human endothelial cells, and such endothelial RANK up-regulation results in enhancing angiogenic activities of TRANCE in endothelial cells.

Retroviral Vectors and Generation of Stable
Transfectants-Dominant-negative MEK1 (DN-MEK1) and p85 (⌬p85) have been characterized (25,26). The cDNA sequences encoding of hemagglutinin (HA)tagged DN-MEK1 and Myc-tagged ⌬p85 were subcloned into pMSCVpuro vector (Clontech, Palo Alto, CA). The vectors were introduced into HEK293T cells (packaging cell line) with 1 g of pVSV-G vector (Clontech) using the LipofectAMINE Plus reagents according to the instructions from the manufacturer. The next day, virus supernatant from these cells added with 5 g/ml Polybrene to HUVECs. After 24 h of incubation, the medium was replaced with fresh medium containing 3 g/ml puromycin. The selection of clones was carried out for 1 week in the presence of 3 g/ml puromycin. Protein expression was confirmed by Western blot.
Semi-quantitative RT-PCR Analysis-Total RNA was obtained from HUVECs using the TRIzol reagent kit. The different amounts of total RNA (0.5-5 g) isolated from HUVECs were subjected to reverse transcriptase-polymerase chain reaction (RT-PCR) assay, and the correlation between the amounts of RNA used and those of PCR products from both target (RANK) and internal standard (␤-actin) mRNAs was examined. Briefly, target RNA was converted to cDNA by treatment with 200 units of reverse transcriptase and 500 ng of oligo(dT) primer in 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl 2 , 10 mM dithiothreitol, and 1 mM dNTPs at 42°C for 1 h. The reaction was stopped by heating at 70°C for 15 min. Three l of the cDNA mixture was used for enzymatic amplification. Polymerase chain reaction was performed in 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl 2 , 0.2 mM dNTPs, 2.5 units of Taq DNA polymerase, and 0.1 M each of primers for RANK. The amplification was performed in a DNA thermal cycler (model PTC-200; MJ Research) under the following condition: denaturation at 94°C for 5 min for the first cycle and for 30 s starting from the second cycle, annealing at 65°C for 30 s, and extension at 72°C for 30 s for 20 repetitive cycles. Final extension was at 72°C for 10 min. The primers used were 5Ј-TTAAGCCAGTGCTTCACGGG-3Ј (sense) and 5Ј-ACGTA-GACCACGATGATGTCGC-3Ј (antisense) for the RANK. The PCR products were electrophoresed on a 1.5% agarose gel and stained with ethidium bromide. The amounts of PCR products were quantitated by densitometry. The 497-bp RANK amplicons generated by RT-PCR from HUVECs were directly sequenced using an ABI Prism cycle sequencing kit (PerkinElmer Life Sciences), and DNA sequences were compared with published sequences to confirm their identity using computation performed at the NCBI and the BLAST network service.
FACScan Analysis-Cells from subconfluent cultures were detached gently from the plate by treating with PBS containing 2 mM EDTA. The cells were washed two or three times with PBS and resuspended in PBS containing 3% bovine serum albumin. Subsequently, the cell suspensions were incubated with FITC-conjugated TRANCE (3 g/ml) for 30 min on ice. The cells were washed two times with PBS and fixed with PBS containing 1% formalin for 30 min on ice. The fluorescence intensity was determined by FACScan analysis.
Tube Formation Assay-Tube formation assay was performed as previously described (27). Briefly, 250 l of growth factor-reduced Matrigel (10 mg of protein/ml) was pipetted into a 16-mm diameter tissue culture well and polymerized for 30 min at 37°C. HUVECs incubated in M199 containing 1% FBS for 6 h were harvested after trypsin treatment, resuspended in M199, plated onto the layer of Matrigel at a density of 1 ϫ 10 5 or 4 ϫ 10 5 cells/well, and followed by the addition of 5 g/ml TRANCE. After 20 h, the cultures were photographed (magnification, ϫ40). The area covered by the tube network was determined using an optical imaging technique in which pictures of the tubes were scanned in Adobe Photoshop and quantitated using Image-Pro Plus (Media Cybernetics).
Western Blotting-Cell lysates from HUVECs were electrophoresed on SDS-PAGE gel and transferred to polyvinylidene difluoride membrane. The blocked membranes were then incubated with the indicated antibody, and the immunoreactive bands were visualized using chemiluminescent reagent as recommended by Amersham Biosciences.
Data Analysis and Statistics-The data are presented as means Ϯ S.E. from three different experiments. Statistical comparisons between groups were performed using Student's t test or analysis of variance in combination with Duncan's multiple range test.

VEGF Increases Expression of RANK mRNA and Protein in
Endothelial Cells-VEGF has been shown to control expression of many genes closely associated with induction and maintenance of the neovasculature in human (28). To examine whether VEGF regulates RANK expression in endothelial cells, subconfluent HUVECs were stimulated with 20 ng/ml VEGF for various time periods. Total RNA was then extracted from the culture, and semi-quantitative RT-PCR was performed for the detection of RANK mRNA expression using specific primers for human RANK with an endogenous standard (␤-actin) for internal normalization. The identity of the amplified fragment (0.497 kb) was verified by sequencing analysis. As shown in Fig. 1A, VEGF markedly increased the RANK transcript in HUVECs; maximal activation was observed 4 h after VEGF stimulation and declined thereafter. We also observed that VEGF enhances RANK mRNA expression in human microvascular endothelial cell line, HMEC-1 (Fig. 1B). To further confirm the effect of VEGF on RANK expression, the surface expression of RANK protein in endothelial cells was assessed by FACS analysis using FITC-conjugated TRANCE. VEGF increased RANK protein level on the surface of HUVECs in a time-dependent manner, reaching a near-maximal level within 8 h (Fig. 1D). The change at 8 h in the expression level of RANK relative to control ranged between 2-and 5-fold; the average -fold increase was about 3. These results indicate that VEGF enhanced RANK mRNA and protein expression in endothelial cells.
VEGF Increases RANK Expression through Its Receptor Flk-1/KDR but Not Flt-1-VEGF exerts its effects through binding to its two receptor tyrosine kinases, Flk-1/KDR and Flt-1, which are expressed on endothelial cells (29 -31). To determine which VEGF receptor is involved in RANK expression, we tested the effects of PlGF. PlGF is a member of the VEGF family of proteins that binds Flt-1 with high affinity but fails to bind Flk-1/KDR (32). The results showed that PIGF had no significant effect on RANK mRNA and protein expression, suggesting that activation of the Flk-1/KDR receptor may be responsible for VEGF-induced RANK expression, whereas Flt-1 is not involved (Fig. 2, A and B). The involvement of KDR receptor in RANK up-regulation was further examined by using SU1498, a specific inhibitor of Flk-1/KDR. Pretreatment of HUVECs with SU1498 to inactivate KDR completely blocked the effect of VEGF on RANK mRNA expression (Fig. 2, A and  B). Thus, these results show that Flk-1/KDR is the dominant receptor involved in VEGF-induced increase in RANK expression.
Src, PLC, PKC, and Phosphatidylinositol 3Ј-Kinase (PI3K) Were Involved in VEGF-stimulated RANK Expression in HUVECs-Many studies have shown the importance of Src protein-tyrosine kinases, PLC, and PI3K in Flk-1/KDR signaling and VEGF-induced endothelial gene expression (33)(34)(35). In particular, activation of c-Src was reported to mediate VEGF signaling through PLC-␥ (36). To address the role of these signaling molecules in VEGF-induced RANK up-regulation, HUVECs were treated in medium containing VEGF, as well as inhibitors to Src tyrosine kinases, PLC, or PI3K. Fig. 3A shows that PP1, a potent Src tyrosine kinase inhibitor, significantly blocked VEGF-induced expression of RANK mRNA. Inhibition of PLC with its specific inhibitor U73122 also blocked VEGFinduced RANK mRNA up-regulation, whereas U73343, a structurally similar derivative of U73122 but inactive for PLC, had no significant effect (Fig. 3A). In addition, the PI3K inhibitors LY 294002 and wortmannin markedly inhibited VEGF-induced increase in RANK mRNA expression (Fig. 3B). FACS analysis further showed that VEGF-induced cell surface expression of RANK protein was also significantly inhibited by these signaling blockers (Fig. 3D). To further evaluate the relevance of PI3K in VEGF-induced RANK expression, HUVECs were either mock-infected or infected with retrovirus encoding ⌬p85, a dominant negative mutant of the p85 regulatory subunit of PI3K that lacks the binding site for the p110 catalytic subunit (26). The expression level of ⌬p85 was confirmed by Western blot analysis (Fig. 4A). RANK expression was induced by VEGF in mock-infected cells. However, overexpression of ⌬p85 significantly ablated both RANK mRNA and surface protein expression in response to VEGF (Fig. 4, B and C). These results suggest that activation of PLC and PI3K was involved in VEGF-induced RANK expression.
PLC is known to generate inositol triphosphate and diacylglycerol, which activate intracellular Ca 2ϩ mobilization and PKC, respectively. To further characterize the role of the downstream signaling pathway following PLC activation in RANK up-regulation, we assessed the abilities of the intracellular Ca 2ϩ chelator BAPTA-AM and the broad PKC inhibitor chelerythrine chloride to block RANK expression by VEGF. Inhibition of PKC with chelerythrine chloride blocked VEGF-induced RANK expression to the basal level. However, BAPTA-AM only had no significant effect (Fig. 3, C and D). These data demonstrate that the VEGF-induced increase in RANK expression is dependent on PKC activity rather than the role of intracellular Ca 2ϩ .
Inhibition of MAPK Blocks VEGF-induced RANK Up-regulation-Activation of MAPK is known to be importantly involved in the VEGF signaling pathway in endothelial cells. The MAPK pathway can be activated via several upstream signaling routes. VEGF-induced MAPK activation is previously re-ported to occur via a PKC-dependent pathway (37). Because VEGF-induced RANK up-regulation was blocked by inhibitors of Src, PLC, and PKC, which are linked to the ERK cascade, we first confirmed the involvement of Src, PLC, and PKC in ERK activation in our system and then evaluated whether activation of ERK was required for VEGF-induced RANK expression. Subconfluent HUVECs were exposed to VEGF with or without inhibitors of Src, PLC, and PKC, and the activation of ERK1/2 (p44 ERK1 and p42 ERK2) was analyzed by Western blot analysis using antibody directed against the phosphorylated form of ERK1/2. As shown in Fig. 5A, inhibitors of Src, PLC, and PKC significantly blocked VEGF-induced ERK activation, indicating that activation of ERKs by VEGF occurs through the Src-PLC-PKC pathway. Therefore, we assessed the role of ERK1/2 in VEGF-induced RANK up-regulation. The effects of VEGF on RANK expression in mRNA and protein levels were completely blocked by the specific MAPK inhibitor PD98059 (Fig. 5, B and C). To further evaluate the relevance of ERKs in VEGF-induced RANK expression, HUVECs were either mockinfected or infected with retrovirus encoding dominant negative MEK1 (DN-MEK1). The expression level of DN-MEK1 was confirmed by Western blot analysis (Fig. 5D). Expression of DN-MEK1 resulted in substantial inhibition of both RANK mRNA and surface protein expression in response to VEGF (Fig. 5, E and F). These results suggest that the MAPK pathway is required for VEGF-induced RANK up-regulation in endothelial cells.
Our data showed that the PI3K inhibitor was effective for suppressing VEGF-induced RANK up-regulation (Fig. 3, B and  D). Previous studies have shown that nitric oxide (NO), reactive oxygen species, and ERK1/2 lie downstream of PI3K in the VEGF signaling pathway. In particular, it is noteworthy that a negative or positive regulatory role of the PI3K signaling pathway in the activation of ERK1/2 has been reported in various different cell types (33,38). Thus, we tested the effects of PI3K inhibitors on the ERK activity in primary cultured HUVECs. Fig. 5A shows that LY294002 significantly inhibited VEGFinduced ERK activation. A similar result was obtained by another PI3K inhibitor wortmannin (data not shown). However, neither the NO synthase inhibitor N G -nitro-L-arginine methyl ester (L-NAME) nor the antioxidant N-acetyl-L-cysteine (NAC) had a significant effect on VEGF-stimulated ERK1/2 activation and RANK up-regulation (data not shown). These results suggest that the PI3K pathway is involved in VEGF-induced RANK expression in endothelial cells partly through activation of ERK1/2.

The Relative Effect of Various Angiogenic Factors on RANK Expression-Angiogenesis in vivo
is modulated by a wide variety of molecules acting on endothelial cells. Thus, the relative effects of several other angiogenic factors including bFGF, HGF, EGF, IL-8, and TGF-␤ on RANK expression were investigated. HUVECs were treated with various growth factors, and RANK mRNA and surface protein expression was analyzed by semi-quantitative RT-PCR and FACS analysis, respectively. Fig. 6 shows that, in addition to VEGF, bFGF and HGF were also capable of increasing both RANK mRNA and surface protein expression, whereas EGF, IL-8, and TGF-␤ had no significant effect. Considering previous observations that bFGF and HGF strongly induce ERK1/2 activation in endothelial cells, such signaling pathway is likely to be responsible for RANK up-regulation by bFGF and HGF. activation in HUVECs via activation of its receptor RANK (22). Because VEGF increased the cell surface expression of RANK in endothelial cells, it is expected that the angiogenic activities and related signaling pathway of TRANCE may be enhanced in RANK up-regulated endothelial cells. Thus, we evaluated the effect of VEGF pretreatment for 8 h on TRANCE-induced ERK1/2 activation and tube formation. To examine the ERK activity, HUVECs were pre-incubated for 8 h with or without VEGF and then stimulated with TRANCE. As shown in Fig.  7A, the effect of TRANCE on the ERK activity was markedly increased by VEGF pretreatment. Because the ERK activity induced by VEGF was reduced to the near basal level at 8 h, the increase of the TRANCE effect on activation of ERK1/2 is not the result of the synergistic effect of VEGF on the ERK pathway. To test the effect of RANK up-regulation on TRANCEinduced tube formation, HUVECs were pre-incubated for 8 h with or without VEGF and then re-plated on growth factorreduced Matrigel. In the absence of angiogenic factors, both control cells devoid of VEGF pretreatment and cells pretreated with VEGF formed incomplete and narrow tube-like structures on Matrigel (Fig. 7B, a and e). In the control cells, TRANCE or VEGF alone increased tube formation by ϳ2-fold on Matrigel The cells were stimulated with VEGF (20 ng/ml) for 4 h. Total mRNAs were isolated and semi-quantitative RT-PCR was performed as described under "Material and Methods." Actin (500 bp) was used as an internal control. In graph, the inset shows results from a representative RT-PCR trial for RANK (top bands) and actin (lower bands in each set) corresponding to the conditions represented by the bars. All data were expressed as the mean Ϯ S.E. percentage of the RANK/actin mRNA expression determined in control untreated HUVECs (set at 100%). Data shown in the bar graph were compiled from at least three independent experiments. C and F, HUVECs or cells expressing DN-MEK1 were treated as the same in B and E except for 8-h stimulation. The cell surface expression of RANK was analyzed by FACScan analysis as described in Fig. 1C. CTL, control untreated cells. Data are mean Ϯ S.E. percentage of the relative RANK protein level determined in control cells (set at 100%) from three different experiments. *, p Ͻ 0.05 versus control; #, p Ͻ 0.05 versus VEGF only; ##, p Ͻ 0.01 versus VEGF only. (Fig. 7B, b and c) but co-stimulation showed no significant synergistic effect (Fig. 7B, d). In cells pretreated with VEGF, TRANCE showed the higher activity to induce the formation of well organized and robust tube-like structures than that in the control cells (Fig. 7B, f), whereas in cells co-incubated with SU1498 to inactivate KDR during VEGF pretreatment this response was abrogated (Fig. 7B, h). All together, these results suggest that VEGF potentiated the angiogenic activities of TRANCE via up-regulation of RANK, the TRANCE receptor, in endothelial cells. DISCUSSION RANK is known to be an essential signaling receptor for TRANCE, which regulates the development of osteoclasts and lymph nodes and also influences the development of the B-cell lineage and dendritic cell function in vivo (39). Recently, it was also reported that TRANCE-RANK engagement plays important roles in mammary gland development and angiogenesis (22,40). Thus, RANK is thought to control a variety of biological phenomena in mammalian cells. RANK mRNA is ubiquitously expressed in mammalian tissues, and the expression of RANK protein was also shown in several cell types such as myeloid-derived dendritic cells, activated T cells, osteoclast progenitors, mammary epithelial cells, and endothelial cells (39). Although it has been reported that TGF-␤ and IL-4 up-regulated RANK expression in peripheral blood T lymphocytes (41) and the murine monocytic cell line, RAW 264.7 (42), the regulatory mechanism of RANK expression has not been well elucidated. In the present study, we demonstrate that VEGF, a potent angiogenic factor, positively regulates RANK expression in endothelial cells. Semi-quantitative RT-PCR and FACS analysis reveal that VEGF can significantly increase both RANK mRNA and surface protein expression in human endothelial cells. Furthermore, we show that other angiogenic factors such as bFGF and HGF can induce RANK mRNA expression in endothelial cells. Thus, our results suggest that various growth factors and cytokines may control RANK expression and such regulation of the cellular RANK level may serve as a regulatory mechanism for the biological action of TRANCE.

Pretreatment of HUVECs with VEGF Potentiates the Effects of TRANCE on ERK Activation and Tube Formation-We
The present study also demonstrates the VEGF-elicited pathways involved in the induction of RANK in endothelial cells (Fig. 8). Our results showing that PIGF failed to induce RANK expression and the KDR inhibitor SU1498 blocked VEGF-mediated RANK up-regulation indicate that VEGF increased RANK expression mainly through its receptor Flk-1/ KDR but not Flt-1. Furthermore, we showed that the PLC-PKC pathway is critical for VEGF-mediated RANK expression. The role of PKC in VEGF-dependent gene expression in endothelial cells has been reported previously. Shen et al. (34) have shown that VEGF treatment induces a rapid redistribution of PKC-␣, -␥, and -⑀ from cytosolic to membrane fractions, indicating that VEGF specifically activates these three PKC isoforms in HUVECs. In addition, they revealed that inhibition of PKC with its specific inhibitors prevents VEGF-induced endothelial nitric-oxide synthase (eNOS) up-regulation in endothelial cells. PKC is known to activate the NF-B and ERK pathway. Indeed, VEGF-induced activation of PLC-␥ and PKC is an essential step for induction of cell adhesion molecule mRNAs such as intercellular adhesion molecule 1, VCAM-1, and E-selectin in endothelial cells, and the induction occurs via activation of NF-B (35). Recently, the PKC-ERK system was also shown to be crucial for VEGF-induced tissue factor production (33). Our data show that inhibition of upstream signaling molecules of PKC such as Src and PLC or direct inhibition of PKC blocked VEGF-induced ERK activation in HUVECs (Fig. 5A). Moreover, the ERK pathway inhibitor PD98059 almost completely abrogated the VEGF-induced RANK expression. Thus, these results imply that the PKC-dependent ERK activation is critical for VEGF-induced RANK expression in endothelial cells. Given the result that VEGF induces activation of transcription factor AP-1 via the ERK cascade in endothelial cells (43), it is suggested that activation of AP-1 may be importantly involved in the induction of RANK transcript by VEGF. More studies will be required to elucidate the specific transcription factors engaged in VEGF-induced RANK expression.
In addition to the role of the PLC-PKC pathway in VEGFinduced RANK up-regulation, our data also show the significant role of PI3K. In contrast to our observation, the negative role of PI3K in VEGF-mediated vascular gene expression has been previously suggested. Blum et al. (33) revealed that inhibition of PI3K with wortmannin or LY294002 resulted in a strong enhancement of the VEGF-induced tissue factor production, indicating a negative regulatory role of the PI3K in tissue factor-inducing pathways. Moreover, the PI3K inhibitor enhanced both basal and VEGF-stimulated adhesion molecule expression, whereas insulin, a PI3K activator, suppressed both basal and VEGF-stimulated expression (35). In marked contrast, our results show that PI3K inhibitors or ⌬p85, a dominant negative form of PI3K, significantly reduced the stimulatory effect of VEGF on RANK expression. By using a cell viability test, we confirmed that this inhibitory activity of PI3K FIG. 6. The relative effect of various angiogenic factors on RANK expression. A, HUVECs were stimulated with VEGF (20 ng/ ml), bFGF (10 ng/ml), HGF (50 ng/ml), EGF (20 ng/ml), IL-8 (100 ng/ml), or TGF-␤ (5 ng/ml) for 4 h. Total mRNAs were isolated, and semi-quantitative RT-PCR was performed as described under "Material and Methods." Actin (500 bp) was used as an internal control. In graph, the inset shows results from a representative RT-PCR trial for RANK (top bands) and actin (lower bands in each set) corresponding to the conditions represented by the bars. All data were expressed as the mean Ϯ S.E. percentage of the RANK/actin mRNA expression determined in control untreated HUVECs (set at 100%). B, HUVECs were treated as the same in A except for 8 h of stimulation. The cell surface expression of RANK was analyzed by FACScan analysis as described in Fig. 1C. CTL, control untreated cells. Data shown in the bar graph were compiled from at least three independent experiments. *, p Ͻ 0.05 versus control.
is not a result of the cytotoxicity (data not shown). To investigate the mechanisms downstream of PI3K, we tested the effect of the NOS inhibitor L-NAME or the antioxidant NAC. However, these inhibitors had no effect on VEGF-induced RANK mRNA expression. Interestingly, we observed that wortmannin or LY294002 partly but significantly inhibited VEGF-induced ERK activation in HUVECs, indicating the positive regulatory role of the PI3K in VEGF-mediated ERK signaling pathway. Although a negative regulatory role of the PI3K in VEGFmediated ERK activation has been proposed in HUVECs (33), many recent observations also support the stimulatory role of PI3K in ERK activation in response to VEGF (38). Thus, the PI3K-dependent ERK activation is likely to be in part involved in VEGF-induced RANK expression. Additionally, our data together with other previous reports suggest the differential role of PI3K in VEGF-elicited endothelial gene expression.
In a previous study, we demonstrated angiogenic activities of TRANCE in vitro and in vivo (22). TRANCE stimulated proliferation, migration, and tube formation of human endothelial cells. Our data showing RANK up-regulation after VEGF treatment prompted us to explore whether increased RANK expression in cultured endothelial cells would potentiate the angiogenic activity of TRANCE. We first examined the effect of VEGF-induced RANK up-regulation on ERK activation by TRANCE. The results showed that in RANK overexpressed HUVECs with VEGF pretreatment, the TRANCE response was more potent in comparison with that in control cells (Fig.  7A). We further performed tube formation assay on two-dimensional Matrigel to examine the effect of RANK induction on TRANCE-induced endothelial cell morphogenesis and revealed that it markedly elevates the TRANCE effect on endothelial cell morphogenetic differentiation (Fig. 7B). These results indicate that the angiogenic activity of TRANCE is modulated by FIG. 8. Schematic illustration of VEGF signaling pathway linked to RANK expression in endothelial cells.

FIG. 7. Pretreatment of HUVECs with VEGF potentiates the effects of TRANCE on ERK activation and tube formation.
A, the effect of VEGF pretreatment on TRANCE-induced ERK activation. HUVECs were preincubated for 8 h with or without VEGF (20 ng/ml) prior to stimulation with TRANCE (3 g/ml) for 20 min. Phosphorylated forms of ERKs (P-ERK1 and P-ERK2) in whole cell extracts were detected with phosphospecific antibody. The membranes were stripped and reprobed with anti-ERK2 antibody. This is a representative result of two independent experiments. B, the effect of VEGF pretreatment on TRANCE-induced tube formation. HUVECs were cultured for 8 h in the absence (a-d) or presence (e-h) of VEGF (20 ng/ml). Cells (g and h) were preincubated for 30 min with 20 M SU1498 prior to VEGF treatment. Then, cells (a-h) were collected and re-plated on Matrigel-coated plates at a density of 4 ϫ 10 5 cells/well without (a, e, and g) or with 3 g/ml TRANCE alone (b, f, and h), 20 ng/ml VEGF alone (c), or 3 g/ml TRANCE plus 20 ng/ml VEGF (d). After 20 h, photographs were taken (magnification, ϫ40). C, the area covered by the tube network was quantitated using Image-Pro Plus software. Data are mean Ϯ S.E. percentage from three different experiments.
VEGF via up-regulation of its surface receptor expression in endothelial cells.
Numerous observations have shown that VEGF is generated from a variety of cells inside or in the close vicinity of blood vessels under many physiological or pathological conditions and it acts as a major player at the initial step of angiogenesis. VEGF not only increases endothelial cell permeability that is prerequisite for cell migration and proliferation but also induces the expression of various vascular endothelial genes such as eNOS, tissue factor, intercellular adhesion molecule 1, Eselectin, and connective tissue growth factor (33)(34)(35)(36). Therefore, it is most likely that endothelial cell RANK expression by VEGF may occur at an early step of neovascularization and endow the specific angiogenic activity of TRANCE toward endothelial cells in a newly growing blood vessel. In this context, it is worthy to note that VEGF generated from hypertrophic chondrocytes in the epiphyseal growth plate is critical for initial blood vessel invasion into cartilage, an avascular tissue, in endochondral bone formation (8). During this process, the vasculature is crucial for paving the way for a variety of cells essential for bone morphogenesis, including osteoblasts, to migrate into cartilage (3,8). Osteoblasts express TRANCE, which can interact with RANK on the surface of endothelial cells during migration into cartilage. We supposed that VEGF promotes initial blood vessel invasion into cartilage and elevates endothelial RANK level in such invading blood vessels; it seems possible that interaction of endothelial RANK with TRANCE expressed on osteoblasts may facilitate the specific recruitment of osteoblasts into cartilage.
Our observations also support the role of VEGF in osteoclast differentiation. Recent studies demonstrated that VEGF could substitute for macrophage colony-stimulating factor in the support of osteoclastic bone resorption (12). VEGF was shown to induce osteoclast differentiation and enhance survival of mature osteoclasts (11). Considering that expression of RANK in osteoclast precursors, which is differentiated from myeloid progenitors by macrophage colony-stimulating factor, is essential for osteoclast maturation, the stimulatory effect of VEGF on osteoclast differentiation may be in part caused by its ability to induce RANK expression in myeloid progenitors or osteoclast precursors. It is also worthy to note that some tumor cells such as breast cancer and prostate cancer induce osteoclastic osteolysis to grow and invade mineralized bone (44). Although the mechanism of tumor-induced osteoclastogenesis is still not understood, the production of direct or indirect osteoclast activating factors from tumor cells is thought to be involved in osteolysis for tumor cell colonization into the bone (44). The bulk of the evidence suggests that tumor cells or accessory cells attracted to the site of neoplastic cell growth generate VEGF when their environment becomes hypoxic or inflammatory (4,45). Thus, the possibility is raised that VEGF produced from tumor tissues may be responsible for tumor-induced osteolysis either by stimulating osteoclast differentiation or enhancing the osteoclastic activity through up-regulation of RANK. These possibilities are under investigation.
The RANK ligand TRANCE was demonstrated to be involved in bone formation and cartilage destruction in a rat adjuvant arthritis model. In addition, it is also shown that TRANCE expression can be regulated in stromal cells by a variety of cytokines generated by tumor-associated macrophages (46) and that the inflammatory cytokines TNF-␣ and IL-1␣ elevate TRANCE expression in human microvascular endothelial cells (47), suggesting its role in the formation of microvessels at a tumor site or an inflammatory site. Indeed, we have previously reported the role of TRANCE in stimulating angiogenesis in vivo and in vitro (22). The present study shows that VEGF, an important angiogenic factor closely associated with induction and maintenance of the neovasculature, increases RANK expression in endothelial cells. Therefore, up-regulation of RANK by VEGF and concomitantly its ligand coupling are likely to contribute to promoting neovessel formation under many physiological and pathological conditions.