Introduction

New blood vessel formation, termed angiogenesis, is an essential process in normal physiology, including tissue development and wound healing, as well as in many pathological conditions such as cancer and diabetes, among other1,2. Endothelial cells play a central role in angiogenesis, and the major driving force for endothelial cell activation is signaling through vascular endothelial growth factors (VEGFs) and their receptors (VEGFRs). Among the VEGF-VEGFR signaling pathways, the VEGF-VEGFR2 axis is the most prominent pathway in angiogenesis. Therefore, targeting this signaling pathway is one of the most promising anti-angiogenic strategies. To establish angiogenic therapies, detailed studies of the molecular mechanisms underlying angiogenesis have been conducted. For example, such studies have led to the development of therapeutic agents such as Avastin and their clinical application. However, the clinical outcomes of angiogenic therapies have not been satisfactory, indicating the need for additional approaches.

The VEGF-VEGFR2 signaling axis remains an important therapeutic target. Most previous studies have focused on the transcriptional and translational regulation of VEGF and VEGFR2. Recently, regulation via post-transcriptional and post-translational mechanisms has gained attention in studies of angiogenesis. Thus, microRNAs such as microRNA-16 have been reported to target VEGFR2 mRNA at the post-transcriptional level3,4, and SCFβ-TRCP has been found to ubiquitinate and degrade VEGFR2 protein5. Furthermore, neddylation6, which involves the conjugation of the ubiquitin-like protein NEDD8 to its target protein, is a crucial post-translational modification in addition to ubiquitination. Neddylation is reportedly required for angiogenic regulation. Importantly, MLN4924, an inhibitor of NEDD8-activating enzyme (NAE), blocks angiogenesis in various models in vitro and in vivo7. MLN4924 is known to inhibit Cullin (CUL)-RING UbE3 ligase activity, which requires CUL protein modification by NEDD8. MLN4924 has also been reported to suppress the survival of several tumors as well as their growth and metastasis. MLN4924 is currently in Phase I clinical trials for the treatment of cancer8,9,10,11,12.

The CUL gene family in humans comprises eight members (CUL1, CUL2, CUL3, CUL4A, CUL4B, CUL5, CUL7 and PARC). The amino acid sequences of these members are highly homologous13. Previous studies have indicated that CULs regulate angiogenesis. For example, CUL2 positively regulates hypoxia-induced angiogenesis by degrading von Hippel-Lindau tumor suppressor (VHL), which is a UbE3 ligase for hypoxia-inducible factor (HIF-1)14. CUL5 regulates angiogenesis through MAPK phosphorylation and actin polymerization15. Furthermore, our recent studies unveiled the crosstalk mechanisms between VEGF and Notch signaling via CUL3-mediated protein degradation pathways. VEGF-A induces the stabilization of BAZF (also known as B-cell chronic lymphocytic leukemia/lymphoma 6 member B, BCL6B) mRNA, resulting in an increase in BAZF protein that mediates the poly-ubiquitination and degradation of RBP-Jκ/CBF-1, a key transcriptional regulator of Notch signaling, with CUL3 UbE3 ligase16,17.

In this study, we found that NEDD8-conjugated CUL3 was essential for the activation of VEGF signal transduction by inducing VEGFR2 expression. We also report that activated CUL3 positively regulated angiogenesis by inducing the expression of VEGFR2 as well as NRP1, NOTCH1 and DLL4.

Results

MLN4924 suppressed VEGFR2 expression

It has been previously been shown that MLN4924, an inhibitor of NEDD8-activating E1 enzyme, suppresses VEGF-A-induced angiogenesis in vivo and in vitro7. Using an aortic ring assay, we also found that VEGF-A-induced angiogenesis was strongly inhibited by MLN4924 (data not shown). To reveal the molecular mechanisms underlying this phenomenon, we first analyzed the effects of MLN4924 on the VEGF signal transduction pathway by examining the activation levels of essential kinases, including VEGFR2, Akt and Erk1/2, in cultured HUVECs. HUVECs were treated with MLN4924 (0.3 μM) or Dimethyl sulfoxide (DMSO) for 72 h before VEGF-A stimulation. Without MLN4924, Erk1/2 and VEGFR2 were strongly phosphorylated by VEGF-A stimulation after 10 min, and then the levels gradually decreased. By contrast, cells pretreated with MLN4924 displayed almost no upregulation of phosphorylation of VEGFR2 (Fig. 1A), whereas the efficiency, phosphorylation per receptor content, of VEGFR2 phosphorylation remained unchanged. VEGF-A also induced phosphorylation of Akt in control cells from 10 to 30 min. In MLN4924-treated cells, VEGF-induced phosphorylation of Akt was scarcely observed, whereas high basal levels of Akt phosphorylation were observed compared with the controls (Fig. 1A). Unexpectedly, MLN4924 treatment significantly reduced the levels of VEGFR2 protein (Fig. 1A).

Figure 1: VEGF-A-induced endothelial activation was required for NAE activity.
figure 1

(A) MLN4924 abrogated VEGF-A-stimulated cell activation of VEGFR2, Akt and Erk1/2. HUVECs were pretreated with MLN4924 (0.3 μM) or DMSO for 72 h. After serum starvation for 12 h, the HUVECs were then stimulated with 50 ng/mL VEGF-A for 10 or 30 min. Total or phosphorylated forms of VEGFR2, Akt and Erk1/2 were detected by immunoblotting. β-tubulin was used as an internal control. The VEGFR2 protein levels and phosphorylated levels of VEGFR2, Akt and Erk were quantified using ImageJ software (right panels). *, p < 0.05. (B) HUVECs were treated with 0.1, 0.3 or 0.6 μM MLN4924 for 72 h. VEGFR2 protein levels in HUVECs were analyzed by Western blotting using an anti-VEGFR2 antibody. The relative intensity of each band was statistically analyzed using ImageJ software and normalized relative to β-actin (loading control). (C) VEGFR2 mRNA in HUVECs treated with 0.1, 0.3 or 0.6 μM MLN4924 for 72 h was measured by qRT-PCR. VEGFR2 mRNA levels were normalized to GAPDH mRNA. ***p < 0.001. (D) HUVECs were pretreated with DMSO or 0.3 μM MLN4924 for 72 h. The cells were stimulated with VEGF-A (50 ng/mL) for 2 h. PTGS2 (left panel) and VCAM1 (right panel) mRNA levels were then measured by qRT-PCR. The relative target mRNA levels were normalized to GAPDH mRNA. ***p < 0.001. The experiments were performed independently 3 times.

Full size image

To clarify the effect of MLN4924 on VEGFR2 more precisely, the effects of different concentrations of MLN4924 on VEGFR2 protein and mRNA levels in HUVECs were examined. After the HUVECs were cultured with 0.1, 0.3 or 0.6 μM MLN4924 for 72 h, Western blotting and qRT-PCR analyses were performed. The results showed that both the levels of VEGFR2 protein and mRNA decreased in a dose-dependent manner (Fig. 1B). This suppression largely occurred at concentrations >0.3 μM MLN4924 (up to 80% inhibition) (Fig. 1C).

VEGF-A induces the expression of several angiogenesis-regulating molecules through VEGFR2, such as PTGS2 (cyclooxygenase-2) and vascular cell adhesion molecule-1 (VCAM1)18. To estimate the effect of VEGF stimulation on MLN4924-treated HUVECs, qRT-PCR assessments for PTGS2 and VCAM1 were performed. After the HUVECs were cultured with 0.3 μM MLN4924 or DMSO for 72 h, the cells were stimulated with VEGF-A for 2 h. PTGS2 and VCAM1 mRNA levels were elevated 5.7- and 3.2-fold by VEGF-A stimulation, respectively (Fig. 1D). However, these inductions were almost completely abrogated to basal levels by MLN4924 treatment (Fig. 1D). These results indicated that the inhibition of VEGF-A-induced activation of endothelial cells by MLN4924 treatment was due to the depletion of VEGFR2 mRNA and the resulting downregulation of VEGFR2 protein.

CUL3 was involved in VEGFR2 mRNA expression and VEGF signaling

CUL UbE3 ligases require modification by NEDD8 for their activation6, and the effect of MLN4924 on VEGFR2 expression and its signaling in HUVECs appears to be caused by CUL inactivation. Therefore, we investigated the involvement of six CULs (CUL1, 2, 3, 4A, 4B and 5) in MLN4924-induced abrogation of VEGFR2 expression; the effects of CUL7 and PARC were not assessed because their expression levels are very low in HUVECs according to values reported in a public database (http://157.82.78.238/refexa/mainsearch.jsp). Mission siRNAs (Sigma-Aldrich) targeting CUL1, 2, 3, 4A or 5, and ON-TARGET plus Smart pool CUL3 or 4B siRNA (Dharmacon) were applied to HUVECs prior to the detection of VEGFR2 protein and mRNA by Western blotting and qRT-PCR, respectively. When the HUVECs were transfected with 20 nM of each CUL siRNA, the production of each CUL mRNA was significantly decreased based on semi-quantitative RT-PCR and qRT-PCR analyses (Fig. S1). Interestingly, CUL3 siRNA alone, but not other CUL siRNAs, significantly suppressed VEGFR2 protein production (Fig. 2A) and mRNA expression (Fig. 2B). VEGFR2 expression levels following siRNA knockdown of CUL1, 2, 3,4A or 5 (Sigma-Aldrich) were normalized to GAPDH levels, whereas VEGFR2 expression levels following siRNA knockdown of CUL3 or 4B (Dharmacon) were normalized to β-actin levels because CUL4B knockdown affected GAPDH mRNA levels. We further confirmed the specificity of CUL3 knockdown on the reduction of VEGFR2 protein (approximately 80%) and mRNA (approximately 70%) levels by using two types of siRNAs for CUL3 (Fig. 2C and D). During CUL3 knockdown, we also performed immunofluorescence staining with an antibody against VEGFR2. As shown in Fig. 2E, compared with control cells, CUL3 knockdown resulted in a marked decrease in the immunofluorescence intensity of VEGFR2 on the CUL3-depleted endothelial cell surface.

Figure 2: CUL3 regulated VEGFR2 mRNA expression in HUVECs.
figure 2

(A) HUVECs were transfected with Mission siRNAs (Sigma-Aldrich) or ON-TARGET plus Smart pool siRNA (Dharmacon) against CONT, CUL1, CUL2, CUL3, CUL4A and CUL5, or CONT, CUL3 and CUL4B, and then cultured for 72 h. The proteins were extracted and separated using 10% SDS-polyacrylamide gels. VEGFR2 proteins were then detected by Western blotting using β-actin as a loading control. (B) VEGFR2 mRNA levels were quantified by qRT-PCR in CONT, CUL1, CUL2, CUL3, CUL4A or CUL5 (Sigma-Aldrich)-, and CONT, CUL3 or CUL4B siRNAs (Dharmacon)-transfected HUVECs, and normalized to GAPDH or β-actin mRNA levels. *p < 0.05. (C) HUVECs were transiently transfected with control siRNA (left lane), CUL3 siRNA#1 (middle lane), and CUL3 siRNA#2 (right lane) (Sigma-Aldrich). The cells were harvested 72 h after transfection, and the proteins were analyzed by Western blotting using anti-VEGFR2 and anti-CUL3 antibodies. Anti-β-actin was used as an internal control. (D) CUL3 siRNA#1 or #2 were transfected into HUVECs and cultured for 72 h. VEGFR2 (left panel) and CUL3 mRNA (right panel) levels were measured by qRT-PCR. ***p < 0.001. (E) HUVECs cultured on gelatin-coated coverslips were transfected with control or CUL3 siRNA. The cells were fixed and stained with anti-VEGFR2 antibody, followed by Alexa 488-conjugated goat anti-rabbit IgG. Scale bar: 50 μM. The experiments were performed independently at least 3 times.

Full size image

To elucidate the role of the CUL3 gene in the VEGFR signaling cascade, we examined the signal transduction pathway of VEGFRs using CUL3-depleted HUVECs. Akt and Erk1/2 protein levels were not affected by CUL3 knockdown, whereas VEGFR-2 levels were significantly decreased, and VEGFR-2 phosphorylation was rarely detected following VEGF-A treatment for 10 or 30 min (Fig. 3A). Similar results were obtained for MLN4924-treated HUVECs. Hence, activation of Akt and Erk1/2 downstream of VEGFR2 was completely inhibited by CUL3 siRNA (Fig. 3A). Interestingly, VEGFR1, which is known as a decoy receptor of VEGFR219, was not affected by CUL3 knockdown (Fig. 3A).

Figure 3: Depletion of CUL3 abrogated VEGF-A-induced endothelial cell functions.
figure 3

(A) Control or CUL3 siRNA-transfected HUVECs were incubated for 72 h. After serum starvation for 12 h, the cells were harvested after stimulation by VEGF-A (50 ng/mL) for 10 or 30 min. The levels of phosphorylated (p-VEGFR2, p-Erk1/2 and p-Akt) and total VEGFR1, VEGFR2, Erk, Akt and CUL3 were determined by Western blot analysis. β-tubulin was used as an internal control. The protein levels of VEGFR2, and the phosphorylated levels of VEGFR2, Akt and Erk, were quantified using ImageJ software (right panels). **p < 0.01. (B) After incubating the control or CUL3 siRNA-treated HUVECs for 72 h, the cells were stimulated with VEGF-A (50 ng/mL) for 2 h. The mRNA levels of PTGS2 (upper panel) and VCAM1 (lower panel) were measured by qRT-PCR. The data were normalized to the mRNA levels of GAPDH. ***p < 0.001. (C) Control or CUL3 siRNA-transfected and Azami-Green-expressing HUVEC monolayers were wounded with a 200-μL pipette tip. After VEGF-A (50 ng/mL) stimulation for 6 or 12 h, fluorescence images were captured using a microscope. (D) Quantification of the motility of CUL3 knockdown cells was conducted by measuring the distance migrated compared with the controls. ***, p < 0.001. (E) Control or CUL3 siRNA-transfected HUVECs were stimulated with VEGF-A (50 ng/ml) for 24 h. After staining with Hoechst, the number of nuclei was counted. ***p < 0.001. The experiments were performed independently at least 3 times.

Full size image

We also investigated whether VEGF-VEGFR2 signaling in CUL3-depleted HUVECs affected VEGF-A-induced pro-angiogenic gene expression. As demonstrated by the qRT-PCR analyses shown in Fig. 3B, induction of the pro-angiogenic genes PTGS2 and VCAM1 by VEGF-A stimulation for 2 h was completely abrogated in CUL3-depleted HUVECs. Furthermore, to evaluate the role of CUL3 in VEGF-mediated endothelial cell function, wound healing and cell proliferation assays were performed. Control or CUL3 siRNA-transfected HUVECs were cultured in the presence or absence of human recombinant VEGF-A (50 ng/mL). As shown in Fig. 3C–E, VEGF-A-induced endothelial cell migration and proliferation were almost completely inhibited by CUL3 depletion compared with their controls. Taken together, our results suggested that NEDD8-conjugated CUL3 was essential for the expression of VEGFR2, but not VEGFR1, and VEGF-A-induced angiogenic responses.

These findings indicated that among the 8 members of the CUL family, CUL3 plays a selective and crucial role in VEGFR2 expression and VEGF signaling in HUVECs.

SPOP, an adaptor protein for CUL3, was required for VEGFR2 expression

CUL3 assembles with Roc1/Rbx1/Hrt1 at the C-terminal region and with a Bric-a-brac/Tramtrack/Broad complex (BTB) domain-containing protein (BTBP) at the N-terminal region to form UbE3 ligase complexes. In these complexes, BTBPs play a crucial role in determining the substrate specificity20. One hundred eighty-three genes have been reported to encode BTBPs in human21. These findings suggest that one of the CUL3-BTBP complexes could target one or more VEGFR2 expression-downregulating factors for ubiquitination and subsequent degradation.

To identify the BTBP(s) responsible for acting as an adaptor for CUL3 and regulating VEGFR2 expression, we selected 9 highly expressed BTBP genes in HUVECs using a public database (http://157.82.78.238/refexa/mainsearch.jsp) and performed mRNA knockdown experiments with siRNAs at a concentration of 20 nM. The Western blotting results for VEGFR2 protein in HUVECs (Fig. 4A and Table 1) revealed a decreased signal intensity of VEGFR2 in speckle-type POZ protein (SPOP)-depleted HUVECs. Monitoring of VEGFR2 protein levels by Western blotting revealed a gradual decrease to a level equal to that observed in cells treated with CUL3 siRNA for 72 h (approximately 70–80%) (Fig. 4B). SPOP siRNA reduced the SPOP mRNA levels by more than 80%. Under this condition, SPOP siRNA also reduced VEGFR2 mRNA levels by approximately 80%, whereas the levels of VEGFR1 mRNA exhibited a relative increased (Fig. 4C). By contrast, in HUVECs overexpressing SPOP (via lentiviral infection), there was a significant increase in the levels of VEGFR2 mRNA (approximately 2- to 2.5-fold) and protein (approximately 3- to 4-fold) (Fig. 4D). Moreover, the interaction between CUL3 and SPOP in HUVECs was analyzed by pull-down assays using Halo-tag technology. Endogenous CUL3 was detected in proteins purified by Halo-tagged SPOP-conjugated resin using Western blotting (Fig. 4E). The 42-kDa band corresponding to SPOP was released from the Halo-tag resin complex by TEV protease, and it was clearly detected in the lane containing SPOP-expressing HUVECs by silver staining and Western blotting with an anti-SPOP antibody. The other specific band was not observed in the control lane. These data suggested that CUL3 associated with SPOP, which positively regulated VEGFR2 expression in HUVECs.

Figure 4: SPOP, one of the BTBPs that associates with CUL3, regulated VEGFR2 expression.
figure 4

(A) HUVECs were transfected with nine selected BTBP siRNAs and cultured for 72 h. After harvesting the HUVECs, the proteins were subjected to SDS-PAGE followed by Western blotting using antibodies against VEGFR2 and β-actin. The upper panel shows the Western blotting images of VEGFR2 and β-actin proteins in the BTBP knockdown HUVECs. The lower panel shows the relative band density values, which were quantified statistically using ImageJ software. Control or CUL3 siRNAs were used as negative and positive controls, respectively. (B) Control, CUL3 or SPOP siRNA-transfected HUVECs were collected after incubation for 0, 1, 2 or 3 days. VEGFR2 was detected by Western blotting. The relative intensity of each band was analyzed statistically using ImageJ software and normalized to β-actin (loading control). (C)VEGFR2 (left panel), VEGFR1 (middle panel) and SPOP (right panel) mRNA levels in control or SPOP siRNA-transfected HUVECs were quantified by qRT-PCR. The target mRNAs levels were normalized to GAPDH mRNA. ***, p < 0.001. (D) HUVECs were infected with Azami-Green- (AG) or Halo-tagged SPOP-expressing lentiviruses at a multiplicity of infection (m.o.i.) of 3. The HUVECs were harvested at 72 h after infection. VEGFR2 mRNA and protein were detected using qRT-PCR (left panel) and Western blot analysis (middle and right panel), respectively. ***p < 0.001. (E) Azami-Green- or Halo-tagged SPOP-expressing HUVECs were treated with NP-40 cell lysis buffer, followed by incubation with Halo-link Resin. SPOP and SPOP-associated proteins were eluted with TEV protease. Total protein and purified proteins were subjected to SDS-PAGE, and Western blotting was performed with anti-CUL3 and SPOP antibodies (upper panel). Input and pull-down materials were visualized by CBB (lower left panel) and silver staining (lower right panel), respectively. The arrowhead indicates SPOP released from HaloLink resin beads. The band intensities were plotted, and the peak areas, shown in gray color, were quantified using ImageJ software. The experiments were performed independently at least 3 times.

Full size image
Table 1 siRNAs used for screening BTBP that regulate VEGFR2.
Full size table

DAXX was a substrate for CUL3-SPOP and regulated VEGFR2 expression

VEGFR2 mRNA expression in HUVECs was suppressed by MLN4924, CUL3 siRNA and SPOP siRNA. These results suggested that the neddylated-CUL3-SPOP complex indirectly regulated VEGFR2 mRNA expression. The molecular mechanism underlying this regulation, however, is still unclear. Therefore, we speculated that a direct substrate of CUL3-SPOP UbE3 ligase would be able to regulate VEGFR2 mRNA expression. Several proteins have been reported to be substrates of CUL3-SPOP UbE3 ligase20. Among them, we analyzed DAXX protein levels under the conditions described in Fig. 5A. The Western blotting results (Fig. 5B) showed that the levels of endogenous DAXX protein increased following SPOP or CUL3 knockdown. This upregulation was completely abrogated by DAXX knockdown (Fig. 5B). In addition, a cycloheximide chase assay revealed that the stability of DAXX protein was dramatically increased by CUL3 depletion (Fig. 5C), and an in vivo ubiquitination assay revealed that CUL3 and SPOP overexpression significantly increased the poly-ubiquitination of DAXX (Fig. 5D). If DAXX indeed plays a crucial role in VEGFR2 mRNA expression, we hypothesized that the knockdown of DAXX would rescue the depletion of VEGFR2 mRNA by SPOP siRNA. To test this hypothesis, rescue experiments were performed. Transfection of SPOP siRNA together with DAXX siRNA resulted in the complete recovery of both the mRNA and protein levels of VEGFR2 (Fig. 5E and F). However, the levels of VEGFR1 protein in SPOP knockdown HUVECs were not affected by co-transfection with DAXX siRNA, although their mRNA levels were slightly upregulated (approximately 1.5-fold) (Fig. 5E and F). In addition, the VEGFR2 protein levels in MLN4924-treated HUVECs were also partially recovered by DAXX knockdown (Fig. 5G). Moreover, to confirm whether DAXX could be a master regulator of VEGFR2 expression, we measured the levels of VEGFR2 in various concentrations of DAXX siRNA-transfected HUVECs. qRT-PCR showed that the VEGFR2 mRNA level was dramatically increased by DAXX siRNA treatment in a dose-dependent manner, and a 25 nM dose resulted in increased levels of VEGFR2 (2.5-fold higher than in the absence of DAXX siRNA; Fig. 5H). In contrast to the VEGFR2 level, the VEGFR1 mRNA level was slightly decreased in a dose-dependent manner. The corresponding protein levels of VEGFR2, VEGFR1 and DAXX were determined by Western blot analysis (Fig. 5I). These results suggest that CUL3-SPOP regulated VEGFR2-mediated endothelial cell functions by controlling the level of DAXX. However, the pathway linking DAXX to VEGFR2 mRNA remains unknown.

Figure 5: DAXX, a substrate of the CUL3-SPOP complex, negatively regulated VEGFR2 expression.
figure 5

(A) Simplified diagram illustrating the components of the CUL3-SPOP-DAXX axis. (B) HUVECs were transfected with 20 nM CUL3 or SPOP siRNA. HUVECs were also transfected with 20 nM of DAXX siRNA alone or together with 20 nM of CUL3 or SPOP siRNA. The cell extracts were collected on the third day after transfection followed by Western blotting with an antibody against DAXX or β-actin. (C) HUVECs were transfected with CONT siRNA or CUL3 siRNA and cultured for 72 h, followed by cycloheximide treatment for 0, 2, 4, 6, 9, and 12 h. After harvesting the HUVECs, the proteins were subjected to SDS-PAGE followed by western blotting with antibodies against DAXX and β-actin. (D) The 293T cells were transfected with 2 μg of several combinations of constructs, as described in (D). The poly-ubiquitinated proteins were purified by TUBE and analyzed with anti-V5-tagged antibody. (E) HUVECs were transfected with 20 nM CUL3 or SPOP siRNA. HUVECs were also transfected with 20 nM of DAXX siRNA alone or together with 20 nM of CUL3 or SPOP siRNA. The amounts of VEGFR2 (left panel) and VEGFR1 mRNA (right panel) were determined by qRT-PCR. ***p < 0.001. (F) HUVECs were treated with 20 nM DAXX siRNA alone or together with 20 nM CUL3 or SPOP, proteins were extracted and VEGFR2 and VEGFR1 were detected by Western blotting. β-actin was used as an internal control. (G) After transfection with control or DAXX siRNA, HUVECs were cultured with DMSO or MLN4924 (0.3 μM) for 72 h. HUVECs were harvested, and VEGFR2 and β-actin proteins were detected by Western blotting. (H) HUVECs were treated with DAXX siRNA at various concentrations (0, 1.56, 3.13, 6.25, 12.5 or 25 nM). VEGFR2 (left panel) and VEGFR1 (right panel) mRNA levels were quantified by RT-PCR. (I) Proteins extracted from DAXX siRNA (0, 1.56, 3.13, 6.25, 12.5 or 25 nM)-transfected HUVECs were subjected to SDS-PAGE, followed by Western blotting using anti-VEGFR1, anti-VEGFR2, anti-β-actin and anti-DAXX antibodies.

Full size image

The CUL3-SPOP-DAXX axis regulated the expression of genes essential for angiogenesis

DAXX has been shown to be involved in the regulation of the cell cycle and apoptosis via the expression of various genes22, suggesting that a wide range of target genes are regulated by the CUL3-SPOP-DAXX axis in HUVECs. To investigate in greater detail whether the CUL3-SPOP-DAXX axis could target other genes, we first performed comparative gene expression profiling of control and CUL3 siRNA-treated HUVECs using a DNA microarray. As shown in Fig. 6A, many genes were up- and downregulated in response to CUL3 knockdown. Among the genes that were downregulated more than two-fold, we discovered three of particular interest: NOTCH1, DLL4 and neuropilin1 (NRP1), which are important in angiogenic signaling. DLL4 is a Notch ligand, and DLL4-Notch1 signals negatively regulate angiogenesis, whereas NRP1 is a binding protein of VEGFs and functions as a co-receptor of VEGFR2. The mRNA levels of NOTCH1, DLL4 and NRP1 were validated by qRT-PCR and estimated to be reduced by approximately 60% in the control in CUL3-depleted HUVECs (Fig. 6A,B). We further analyzed whether the expression of these genes was also regulated by the CUL3-SPOP-DAXX axis. qRT-PCR analyses were performed, and the directional fold-change of each gene was confirmed. Significant decreases in NOTCH1, DLL4 and NRP1 were observed in both CUL3 and SPOP knockdown HUVECs (Fig. 6C). In addition, we observed that the suppression of each gene by SPOP knockdown was largely reversed by transfection of DAXX siRNA (Fig. 6C). As shown in Fig. 6D, the corresponding changes in proteins were confirmed. Taken together, these findings strongly suggest that the CUL3-SPOP-DAXX axis regulates angiogenesis by upregulating not only VEGF-VEGFR2-NRP1 but also DLL4-Notch 1 signaling (Fig. 6E).

Figure 6: The CUL3-SPOP-DAXX axis positively regulated NOTCH1, NRP1 and DLL4 expression.
figure 6

(A) DNA microarray analyses (log2 signal intensity) of control siRNA compared with CUL3 siRNA. (B) Five genes were selected from the microarray data. The relative microarray expression levels are shown; fold-change ratios for the microarray analysis were calculated considering the expression levels of GAPDH. (C) HUVECs were cultured with 20 nM CUL3 or SPOP siRNA alone. HUVECs were also cultured with 20 nM SPOP together with DAXX siRNA. NOTCH1 (left panel), DLL4 (middle panel) and NRP1 (right panel) mRNA levels were quantified by RT-PCR. ***p < 0.001. (D) HUVECs were treated with 20 nM control, CUL3 or SPOP siRNA alone or SPOP together with DAXX siRNA. Proteins were extracted from the transfected HUVECs, and NOTCH1, DLL4 and NRP1 were detected by Western blotting. (E) Schematic of the molecular mechanism. DAXX suppresses pro-angiogenic factors, including VEGFR2, NOTCH1, DLL4 and NRP1. When DAXX is degraded by the CUL3-SPOP complex, the expression of pro-angiogenic genes is upregulated, resulting in the progression of angiogenesis.

Full size image

In the present study, the molecular mechanisms by which the CUL3-SPOP-DAXX axis positively regulates VEGFR2 mRNA remain unclear. Therefore, to assess the role of the neddylated-CUL3-SPOP-DAXX axis on VEGFR2 mRNA metabolism, luciferase promoter assays and actinomycin D chase assays were performed because both transcriptional and post-transcriptional regulations are essential for VEGFR2 gene expression. As shown in Fig. S2, however, there were no significant decreases in the VEGFR2 promoter activity when HUVECs were treated with CUL3 siRNA or SPOP siRNA (Fig. S2). In addition, the stability of VEGFR2 mRNA was only slightly decreased by CUL3 and SPOP knockdown (Fig. S3). Similar results were obtained for NRP1, NOTCH1 and DLL4 mRNAs (Fig S3). Based on these findings, it can be speculated that the neddylated-CUL3-SPOP-DAXX axis may regulate pro-angiogenic genes such as VEGFR2 via epigenetic regulation. To test this hypothesis, we harvested CUL3 knockdown, SPOP knockdown and SPOP knockdown together with DAXX knockdown HUVECs. We investigated major histone modifications (methylation of histone H3 at lysine 4 (H3K4me3), methylation of histone H3 at lysine 27 (H3K27me3) and acetylation of histone H3 at lysine 9 (H3K9Ac)) in the promoter regions of target genes by chromatin immunoprecipitation assays. Herein, we provide the first evidence that the chromatin status of these promoters was monovalent (H3K4me3) because of the weak enrichment for H3K27me3 similar to the negative controls, compared with H3K4me3 and H3K9Ac (data not shown). Among the histone modifications, we found that trimethylated forms of histone H3 at lysine 4 (H3K4me3; associated with active chromatin and gene expression) were dramatically reduced by CUL3 and SPOP knockdown in the VEGFR2, DLL4, Nrp1, and Notch1, but not VEGFR1, genes (Fig. S4). These histone modification changes were correlated with their mRNA expression levels (Figs 5E and 6C). However, knockdown of additional DAXX did not rescue the depletion of trimethylated forms in target genes by SPOP siRNA. These results suggest that the CUL3-SPOP axis targets a critical regulator of angiogenic gene expression via histone modification at H3K4me3, and the epigenetic regulator and DAXX may regulate, in a complementary manner, the mRNA levels of angiogenic factors such as VEGFR2, DLL4, Nrp1, and Notch1.

Discussion

The VEGFR2 signaling pathway is important for both vasculogenesis and angiogenesis. Genetic deletion of VEGFR2 from mice results in a complete absence of blood vessels, resulting in embryonic lethality in haploinsufficient animals23,24. Moreover, a large number of studies have demonstrated that pharmacological blockage of VEGF-VEGFR2 signaling can suppress tumor angiogenesis, resulting in decreased tumor growth in vivo. However, a significant number of such patients do not respond to anti-VEGF therapy when used either as a single agent or in combination with chemotherapy; the benefit is relatively short-lived, and the majority of patients relapse and progress25. The vasculature of tumors is abnormal and dysfunctional, displaying dilatation, tortuousness and leakiness, which are characteristics of an imbalance between pro- and anti-angiogenic signals26.

Consequently, it is essential to elucidate the regulatory mechanisms underlying the expression and activation of VEGFR2 and to establish more promising anti-angiogenic therapies. In the present study, we provide the first evidence that a neddylation inhibitor (MLN4924) blocks VEGF-A-induced activation of kinases and gene expression by inactivating CUL3-mediated VEGFR2 expression. These findings are completely correlated with the anti-angiogenic phenotypes related to MLN4924 treatment7 or CUL3 knockdown16. However, we found a discrepancy between MLN4924- and CUL3 siRNA-treated HUVECs in terms of VEGF-A-induced Akt phosphorylation. The basal levels of Akt phosphorylation were increased in MLN-4924-treated HUVECs. A recent study has revealed a large number of NEDD8-target molecules related to cell survival using a proteomic approach in A375 melanoma cells27. Furthermore, phosphorylated levels of Akt were increased in response to MLN4924 treatment in MCF7 and HeLa cell28. By contrast, in HCT116 cells, phosphorylation was decreased. These reports and our data suggest that MLN4924 may modulate Akt phosphorylation levels via an unknown cell type-specific molecular mechanism, excluding CUL3-mediated signaling.

It is known that CUL3 associates with BTBPs that determine substrate specificity. Several CUL3-BTBP-based E3 ligase complexes have been reported to be involved in various events. For example, CUL3-Keap1-Nrf2 is essential for oxidant regulation29, CUL3-BAZF-CBF1 is important for angiogenic switching16 and CUL3-SPOP-Gli1/Gli2 participates in morphological determination during developmen30,31,32. Among the BTBPs, SPOP has been relatively well characterized as a key adaptor protein for CUL3-based ubiquitin ligases. In the present study, we found that knockdown of SPOP down-regulated VEGFR2 mRNA and protein levels (Fig. 4) and that endogenous CUL3 interacted with SPOP in HUVECs. These experimental outcomes of SPOP knockdown were blocked by knocking down DAXX. However, VEGFR2 down-regulation by MLN4924 was not completely rescued by DAXX knockdown. Recent evidence has shown that gene-interacting proteins, such as MLX, EID1, KLF5, ORC6L, MAGEA6, MORF4L2, MRFAP1, MORF4L1, and TAX1BP1, are NEDD8 target protein27, indicating that MLN4924 broadly modulates transcriptional signaling, targeting CUL3-independent post-transcriptional pathways. Therefore, the identification of other target molecule(s) of NEDD8 with regulatory activity towards the angiogenic expression of genes, such as VEGFR2 will be need as a future study. These results suggest that among the downstream NEDD8 signaling pathways, CUL3-SPOP-DAXX is a novel regulatory axis in VEGFR2 gene expression. We also showed that the CUL3-SPOP-DAXX axis was necessary for NRP1, DLL4 and NOTCH1, but not VEGFR1, gene expression.

NRP1 binds to VEGF-A and acts as a co-receptor of VEGFRs33. The signal outputs of the VEGF-A-VEGFR2 pathway are enhanced by the interaction with NRP1. However, DLL4-Notch 1 signaling is also critical for vascular sprouting and stabilization34,35. During angiogenesis, DLL4 in tip cells activates Notch signaling in stalk cells, suppressing the expression of tip cell-specific genes such as DLL4 and VEGFR2. A recent investigation showed that during retinal angiogenesis, inhibition of Notch signaling by a specific inhibitor of γ-secretase DAPT or genetic deletion of DLL4, led to hyper-vascularization due to upregulation of VEGF-VEGFR2 signaling36. These results indicated that the collaboration between VEGF and Notch signaling is essential for angiogenic sprouting. The data presented herein showed that the CUL3-SPOP-DAXX axis could function as a crucial angiogenic mediator through the regulated expression of VEGFR2, NRP1, NOTCH1 and DLL4.

However, the CUL3-SPOP-DAXX axis had no effect on VEGFR1 expression. VEGFR1 is highly expressed on endothelial cells and associates with VEGF-A, VEGF-B, and placenta growth facto37. A VEGFR1 deficiency results in early embryonic lethality characterized by vascular abnormalities with overgrowth phenotypes of endothelial cells23,38. Moreover, the binding affinity of VEGFR1 for VEGF-A is higher than that of VEGFR2, but its tyrosine kinase activity is weaker than that of VEGFR2. Therefore, VEGFR1 has been proposed to negatively modulate angiogenesis by acting as a decoy receptor of VEGFR2.

These findings suggest that the regulation of pro-angiogenic VEGF-VEGFR2-NRP1 and anti-angiogenic DLL4-Notch1 can be coupled, but VEGF-VEGFR2-NRP1 and VEGF-VEGFR1 signaling are functionally distinct and differentially regulated. Based on our data, we propose that the CUL3-SPOP-DAXX axis can positively and effectively regulate angiogenesis by upregulating and balancing VEGF-VEGFR2-NRP1 and DLL4-Notch 1 signaling.

An important question in this study concerned how DAXX could regulate the mRNAs of the above-described angiogenic regulators. DAXX was first identified as a Fas-binding cytosolic protein that regulates cellular apoptosis22. DAXX also localizes in the nucleus and acts as a transcriptional and cell cycle regulator by associating with the PML nuclear body and the nuclear heterochromatin region. DAXX carries the SPOP-binding motif 608 VSSTT612. NEDD8 modification of CUL3 is essential for the degradation of DAXX in 293T cells39,40. In the present study, the endogenous level of DAXX protein was upregulated by knockdown of CUL3 or SPOP in HUVECs (Fig. 5B). We also provided evidence that DAXX negatively regulated VEGFR2 expression. A previous study has demonstrated that DAXX can repress Ets-1, which regulates the expression of angiogenic genes, such as matrix metalloproteinases, VEGFR1 and VEGFR2, by binding to unique sequences (GGAA) on their promoter41. Here, we showed that the CUL3-SPOP-DAXX axis positively regulated the expression of VEGFR2 but not VEGFR1. In addition, luciferase promoter assays revealed that CUL3 and SPOP knockdown did not affect VEGFR2 promoter activity (Fig. S2). Based on these and the previously reported data, it seems unlikely that the CUL3-SPOP-DAXX axis regulates VEGFR2 expression via a transcriptional factor such as Ets-1 in the −780 to + 268 of VEGFR2 promoter region. However, it is possible that the CUL3-SPOP-DAXX axis targets a region upstream of VEGFR2 genes that includes the enhancer region. Although the construct (pGL3-VEGFR2-780, Addgene) used for the luciferase promoter assay contains important cis elements, such as several Sp1 sites, two AP-2 consensus sites and an E-box, the association with DAXX occurs in a region further upstream of the VEGFR2 gene that remains to be elucidated. Further analyses using ChIP sequencing with an anti-DAXX antibody may be needed. Actinomycin D chase assays also demonstrated that the stability of VEGFR2 mRNA was not significantly reduced by CUL3 and SPOP depletion (Fig. S3). These data suggest that post-transcriptional regulation is also not required for the gene expression of VEGFR2 by CUL3-SPOP-DAXX. On the other hand, DAXX is also known to control transcription via epigenetic regulatio42. Several studies have reported that DAXX directly and indirectly regulates gene expression by interacting with various chromatin remodeling proteins such as histone deacetylase, DNA methyl transferase and histone-acetyltransferase CBP/p30043. Furthermore, recent studies have shown that DAXX acts as an H3.3-specific histone chaperone in mammalian cells44. Based on these findings, we investigated the main histone modifications of the promoter regions in the VEGFR2, VEGFR1, NOTCH1, DLL4 and NRP1 genes. We found that the H3K4me3 levels in 4 of the 5 promoters (excluding VEGFR1) were dramatically decreased in response to CUL3 or SPOP depletion; these changes were not rescued by additional DAXX knockdown (Fig. S4). Our results suggest at least two possible mechanisms of CUL3-mediated angiogenic gene induction. First, DAXX is a main VEGFR2 mRNA regulator as a substrate of the CUL3-SPOP axis and may negatively regulate the mRNA levels of angiogenic genes via other types of epigenetic regulation such as microRNAs. Second, the CUL3-SPOP axis may target negative regulators of H3K4 trimethylation for ubiquitination and degradation. These substrates may be responsible for the complementary regulation of angiogenesis-related gene expression. Future experiments using genome-wide ChIP-Seq and microarray analyses are needed to clarify the complete molecular mechanism that regulates angiogenic gene expression. In addition, we also need showing the relevance of CUL3-SPOP-DAXX to angiogenesis through in vivo studies using gene-deficient mice. We believe that our data contribute to our understanding of the precise mechanism by which the CUL3-SPOP-DAXX axis regulates angiogenesis in humans and provide insights that may help to establish a new strategy for the treatment of angiogenesis-associated diseases.

Materials and Methods

Reagents and antibodies

Recombinant human VEGF-A was purchased from R&D Systems (Minneapolis, MN). MLN4924 was from Funakoshi Chemical Co. (Tokyo, Japan). Anti-flk1/VEGFR2 (C-1158: sc-504) and anti-SPOP (C-14: sc-66649) antibody were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-CUL3 (Clone CUL3-9, Catalog# SAB4200180), mouse monoclonal anti-β-actin antibody (clone AC-15) and anti–β-tubulin antibody (clone JDR.3B8) were from Sigma-Aldrich (St. Louis, MO). Anti–phospho-VEGFR2 (Tyr1175)(19A10; #2478), anti–phospho-Akt (Ser473; #9270), anti–Akt (#9272), anti–phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204; #9101), anti–Erk (p44/42 MAPK (Erk1/2); #9102), anti–DAXX (25C12; #4533), anti–VEGF receptor1 (#2893) and anti–Neuropilin1 (D62C6) antibody were from Cell Signaling Technology (Danvers, MA). Anti-HaloTag monoclonal antibody (G9211) was from Promega (Madison, WI). The Wako silver staining kit was obtained from Wako Biochemicals (Osaka, Japan).

Cell culture

Human umbilical vein endothelial cells (HUVECs) were obtained from Cell System (Kirkland, WA). HUVECs were grown in the endothelial growth medium EGM-2 (Lonza, Walkersville, MD). All experiments were performed with HUVECs at passages 2–4.

Lentiviral vector construction

Lentiviral vectors were constructed by inserting cDNAs encoding Halo-tagged SPOP into the lentiviral expression vector CSII-CMV-MCS-IRES2-Bsd. Lentiviral vectors were produced in 293T cells. The expression plasmid for Halo-tagged SPOP (Flexi ORF Clone) was purchased from Promega. Lentiviral expression and packaging vectors were kindly provided by Dr. Miyoshi (RIKEN BioResource Center, Tsukuba, Japan).

RNA interference

The siRNAs were obtained from Sigma-Aldrich and Dharmacon (Lafayette, CO) (listed in Supplementary Table 1). MISSION siRNA Universal Negative Control SIC-001 (Sigma-Aldrich) and siGENOME Non-Targeting siRNA (Dharmacon, D-001206-14-20) were used as control siRNAs. The siRNA transfection was performed using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA) at 20 nM according to the manufacturer’s protocol.

Western blotting

Twenty micrograms of protein from HUVECs were electrophoresed in sodium dodecyl sulfate-polyacrylamide gels electrophoresis (SDS-PAGE). The separated proteins were then transferred to polyvinylidene difluoride membranes that were then blocked with 5% nonfat milk in 0.05% Tween 20/PBS (−) (PBS-T) for 30 min, followed by incubation with a primary antibody (1:1000 [v/v]). After washing with PBS-T, the membrane was labeled with the appropriate horseradish peroxidase-conjugated IgG antibodies (Promega; 1:4000 [v/v]). The proteins were visualized using enhanced chemiluminescence and imaged using a LAS-4000 Image Reader (Fujifilm, Tokyo, Japan).

Isolation of RNA, cDNA library synthesis, and quantitative RT-PCR

Total RNAs were extracted from cells using ISOGEN II (Nippon Gene, Tokyo, Japan) according to the manufacturer’s protocol. One microgram of RNA was used for cDNA synthesis using High Capacity RNA-to-cDNA Master Mix (Applied Biosystems, Foster City, CA). Real-time PCR was carried out (FastStart Universal SYBR Green Master ROX; Roche Diagnostics, Basel, Switzerland) with the ABI 7300/7500 Real-Time PCR system (Applied Biosystems). The sequences of the primers are listed in Supplementary Table 2.

Fluorescence immunostaining

HUVECs were plated on gelatin-coated coverslips and then cultured with EGM-2 for 24 h. siRNA targeting human CUL3 or control siRNA were transiently transfected into the HUVECs, followed by culturing for 72 h. The cells were fixed with a 4% paraformaldehyde solution for 30 min and washed three times with PBS. Triton X-100–permeabilized cells were blocked with PBS containing 4% bovine serum albumin (4% bovine serum albumin-PBS), and the cells were labeled with anti-VEGFR2 antibody (1:1000 dilution) in 4% bovine serum albumin-PBS overnight at room temperature. The cells were then washed three times with PBS and probed with Alexa Fluor 488-conjugated goat anti-rabbit IgG antibody (1:1000 dilution) (Invitrogen) in 4% bovine serum albumin-PBS for 1 h at room temperature. After washing three times with PBS, fluorescent images were obtained using a confocal laser microscope A1R (Nikon Corp, Tokyo, Japan).

Cell migration assay

Azami-Green-expressing HUVECs were seeded onto 35-mm dishes. On the following day, control or CUL3 siRNA was transfected into the cells and cultured for an additional 2 days. After serum starvation in endothelial cell basal medium-2 (EBM-2 containing 0.15% FBS) for 12 h, the cell culture was scratched with a 200-μL pipette tip, followed by incubation with EBM-2 containing VEGF-A or BSA for 12 h. Fluorescence images were obtained at 6-h interval using a microscope (Olympus, Tokyo, Japan).

Cell proliferation assay

HUVECs (105 cells/well) were cultured in a 6-well plate in EGM2 for 24 h. After transfection with siRNA against the control or CUL3, the HUVECs were incubated for 48 h, and then the medium was replaced with EBM2 containing 0.15% FCS for serum starvation. The HUVECs were cultured with 50 ng/mL of VEGF-A for 24 h, and then the nuclei were stained with Hoechst. Fluorescence images were captured using microscopy, and the signals were counted using Image Pro-Plus software (Media Cybernetics, Silver Spring, MD).

Pull-down assay

Halo-tagged SPOP or Azami-Green-expressing HUVECs were lysed in cell extraction buffer (50 mM Tris-HCl pH7.5, 0.15 M NaCl, 0.5% NP40). The samples were sonicated three times for 30 sec. The extracts were incubated with HaloLink Resin (Promega) at 4 °C for 3 h, followed by elution of SPOP and SPOP-binding proteins using TEV protease. The pull-down material was separated by SDS-PAGE, assessed by Western blotting with anti-CUL3, anti-Halo-tag and anti-SPOP antibodies, and visualized by silver staining.

In vivo ubiquitination assay

The 293T cells were transfected with several combinations of constructs (pcDNA3.1), such as FLAG-tagged CUL3, SPOP, V5 tagged DAXX, AGIA-tagged45 Rbx1 and HA-tagged ubiquitin. After culturing the cells for 48 h, MG132 (final concentration, 20 μM) was applied for 8 h to inhibit the function of the proteasome. The cells were lysed with NP40 buffer cell extraction buffer (50 mM Tris-HCl pH 7.5, 0.15 M NaCl, 0.5% NP40) and incubated with Tandem Ubiquitin Binding Entity (TUBE) (Nacalai Tesque, Kyoto, Japan). After the beads were washed with lysis buffer, the binding proteins were eluted with SDS sample buffer. Western blot analyses were performed to detect the V5-tag as described.

Cycloheximide chase assay

HUVECs were treated with CONT or CUL3 siRNA and then incubated for 72 h. After treatment with 25 μg/mL cycloheximide (Wako, Osaka, Japan), the cells were lysed at various time points (0, 2, 4, 6, 9, and 12 h). Western blot analyses for DAXX were performed as described.

Microarray analysis

Total RNA was extracted from control or CUL3 siRNA-transfected HUVECs with Isogen II. We used 500 ng of total RNA to generate double-stranded cDNA. The cDNA was transcribed with DIG-labeled nucleotides (Roche), fragmented, and hybridized to a Gene Chip Human Gene 1.0 ST Array (Affymetrix, Santa Clara, CA) according to the manufacturer’s instructions.

Statistical analysis

Data were based on a minimum of 3 independent experiments. The results are represented as the means ± SEs. Two groups were compared using Student’s t test. Differences were considered significant if the p-values were less than 0.05. Statistical analyses were performed using GraphPad Prism (GraphPad Software, San Diego, CA).

Additional Information

How to cite this article: Sakaue, T. et al. The CUL3-SPOP-DAXX axis is a novel regulator of VEGFR2 expression in vascular endothelial cells. Sci. Rep. 7, 42845; doi: 10.1038/srep42845 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.