SUMOylation Fine-Tunes Endothelial HEY1 in the Regulation of Angiogenesis

Abstract

Background: Angiogenesis, which plays a critical role in embryonic development and tissue repair, is controlled by a set of angiogenic signaling pathways. As a TF (transcription factor) belonging to the basic helix-loop-helix family, HEY (hairy/enhancer of split related with YRPW motif)-1 (YRPW motif, abbreviation of 4 highly conserved amino acids in the motif) has been identified as a key player in developmental angiogenesis. However, the precise mechanisms underlying HEY1's actions in angiogenesis remain largely unknown. Our previous studies have suggested a potential role for posttranslational SUMOylation in the dynamic regulation of vascular development and organization.

Methods: Immunoprecipitation, mass spectrometry, and bioinformatics analysis were used to determine the biochemical characteristics of HEY1 SUMOylation. The promoter-binding capability of HEY1 was determined by chromatin immunoprecipitation, dual luciferase, and electrophoretic mobility shift assays. The dimerization pattern of HEY1 was determined by coimmunoprecipitation. The angiogenic capabilities of endothelial cells were assessed by CCK-8 (cell counting kit-8), 5-ethynyl-2-deoxyuridine staining, wound healing, transwell, and sprouting assays. Embryonic and postnatal vascular growth in mouse tissues, matrigel plug assay, cutaneous wound healing model, oxygen-induced retinopathy model, and tumor angiogenesis model were used to investigate the angiogenesis in vivo.

Results: We identified intrinsic endothelial HEY1 SUMOylation at conserved lysines by TRIM28 (tripartite motif containing 28) as the unique E3 ligase. Functionally, SUMOylation facilitated HEY1-mediated suppression of angiogenic RTK (receptor tyrosine kinase) signaling and angiogenesis in primary human endothelial cells and mice with endothelial cell-specific expression of wild-type HEY1 or a SUMOylation-deficient HEY1 mutant. Mechanistically, SUMOylation facilitates HEY1 homodimer formation, which in turn preserves HEY1's DNA-binding capability via recognition of E-box promoter elements. Therefore, SUMOylation maintains HEY1's function as a repressive TF controlling numerous angiogenic genes, including RTKs and Notch pathway components. Proangiogenic stimuli induce HEY1 deSUMOylation, leading to heterodimerization of HEY1 with HES (hairy and enhancer of split)-1, which results in ineffective DNA binding and loss of HEY1's angiogenesis-suppressive activity.

Conclusions: Our findings demonstrate that reversible HEY1 SUMOylation is a molecular mechanism that coordinates endothelial angiogenic signaling and angiogenesis, both in physiological and pathological milieus, by fine-tuning the transcriptional activity of HEY1. Specifically, SUMOylation facilitates the formation of the HEY1 transcriptional complex and enhances its DNA-binding capability in endothelial cells.

Keywords: angiogenesis; retinal diseases; signal transduction; sumoylation; transcription factors.

Figure 1.. HEY1 can be SUMOylated in endothelial cells.

A and B, SUMO modification of HEY1 in 293T cells transfected with Flag-HEY1 and HA-SUMO1 plasmids. C and D Endogenous HEY1 interacts with SUMO1 in HUVECs (C) and MLECs (D). E, Identification of HEY1 SUMOylation sites. Relative HEY1 SUMOylation was normalized to that of input Flag-HEY1 and the WT group. F, Pattern diagram indicating the distribution of the consensus SUMOylation sites in HEY1. G, Detection of the interaction between SUMO1 and HEY1-WT/3KR in 293T cells. H, Detection of the exogenous HEY1-TRIM28 interaction in 293T cells. I, SUMOylation of HEY1 upon overexpression of UBC9 and TRIM28 in 293T cells. Relative HEY1 SUMOylation was normalized to that of input Flag-HEY1 and the “vector + Flag-HEY1” group. J, Detection of the HEY1-TRIM28 interaction in HUVECs. K, SUMOylation of HEY1 upon TRIM28 knockdown in HUVECs. Relative HEY1 SUMOylation was normalized to that of input HEY1 and the si-Ctrl group. L, SUMOylation of HEY1 under hypoxic conditions. HUVECs were treated under normoxic or hypoxic conditions for 48 h. Relative HEY1 SUMOylation was normalized to that of input HEY1 and the normoxia group. M, SUMOylation of HEY1 upon VEGF-A treatment. HUVECs were starved in basic EBM-2 medium for 12 h before treatment with VEGF-A (10 ng/mL) for 15 min. Relative HEY1 SUMOylation was normalized to that of input HEY1 and the no-VEGF-A group. N, Schematic model for the dynamic SUMOylation and de-SUMOylation of HEY1. **P < 0.01, ***P < 0.001 and ****P < 0.0001 (means±SDs; n=6) analyzed by one-way ANOVA with Dunnett's multiple comparisons test (E), two-way ANOVA with Sidak's multiple comparisons test (I) or Student’s t test (K-M). The exact P values are presented in the “Supplemental Statistical Table”.

Figure 2.. SUMOylation facilitates HEY1-mediated repression of angiogenic activity in HUVECs.

The angiogenic capability of HUVECs transduced with Ad-LacZ (as control), Ad-HEY1-WT or Ad-HEY1-3KR vectors was evaluated. A and B, Cell proliferation measured by the Cell Counting Kit-8 (A) and the EdU assay (B). Relative cell proliferation of each group was normalized to that of Day1. C and D, Cell migration measured by the wound healing (C) and the transwell assay (D). E and F, Angiogenic function measured by the sprouting (E) and the tube formation assay (F). Scale bar, 200 μm. G, Gene set enrichment analysis (GSEA) was performed to identify differentially affected functionally related gene sets in HUVECs transduced with Ad-HEY1-WT or Ad-HEY1-3KR vectors. H, RNA-seq heat map of representative genes differentially expressed in HUVECs transduced with Ad-LacZ (as control), Ad-HEY1-WT or Ad-HEY1-3KR vectors. I-K, Validation of downregulation of selected genes upon expression of HEY1-WT in HUVECs by quantitative PCR (I) and immunoblotting (J and K). The mRNA/protein level was normalized to that of β-ACTIN and the control group. *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 (means±SDs; n=6) analyzed by one-way ANOVA with Tukey's multiple comparisons test (B-F, I and K) or two-way ANOVA with Tukey's multiple comparisons test (A) (ns, not significant). Benjamini-Hochberg adjustment was employed for multiple testing correction in I and K, in which adjusted P (false discovery rate corrected Q) values are shown in the “Supplemental Statistical Table”. The exact P values are presented in the “Supplemental Statistical Table”.

Figure 3.. SUMOylation is required for HEY1-mediated inhibition of receptor tyrosine kinase signaling cascades in endothelial cells.

A and B, Immunoblots showing VEGFA-VEGFR2 (A) and bFGF-FGFR1 (B) signaling in HMVECs with HEY1-WT or HEY1-3KR expression by adenoviral delivery. C Immunoblots showing VEGFC-VEGFR3 signaling in HDLECs with HEY1-WT or HEY1-3KR expression by adenoviral delivery. The protein level was normalized to that of β-ACTIN and the “0 min” control group. ***P < 0.001 and ****P < 0.0001 (means±SDs; n=6) analyzed by two-way ANOVA with Tukey's multiple comparisons test (ns, not significant). The exact P values are presented in the “Supplemental Statistical Table”.

Figure 4.. SUMOylation of HEY1 promotes its promoter-binding capability.

A, Differently enriched peaks between HEY1-WT and HEY1-3KR from ChIP-seq data analysis. B, Location of HEY1-WT and HEY1-3KR binding events in relation to closest annotated TSS. C, Enrichment profile of HEY1-WT and HEY1-3KR in relation to closest annotated TSS. D, Location of HEY1-WT and HEY1-3KR binding events respective to various genomic features. E, IGV snapshots showing HEY1-WT ChIP-seq signals at binding sites associated with the VEGFR2, VEGFR3 and FGFR1 promoters in HUVECs. F, ChIP-qPCR results in HUVECs showing the recruitment of HEY1-WT/3KR to the promoters of VEGFR2, VEGFR3 and FGFR1 in HUVECs. The relative promoter enrichment was normalized to that of the input and of the control group. G, Dual luciferase assay showing HEY1-WT/3KR-mediated repression of the VEGFR2 promoter. The relative luciferase activity was normalized to that of Renilla and the control group. H, HEY1-WT/3KR binding motifs identified by HOMER is enriched under ChIP-seq peaks. I, Numbers of E-box/N-box motifs identified in the VEGFR2, FGFR1 or VEGFR3 promoter regions. J, EMSA showing HEY1-WT/3KR binding to E-box and N-box motifs in HUVECs. The relative E/N-box binding was normalized to that of the HEY1-WT group. K, Dual luciferase assay indicating the repressive activity of HEY1-WT/3KR toward E-box or N-box motifs. The relative luciferase activity was normalized to that of Renilla and the control group. L, Dual luciferase assay indicating the repressive activity of HEY1-WT toward wild type E-box or E-box mutant. The relative luciferase activity was normalized to that of Renilla and the WT E-box group. (For ChIP-seq, ChIP-qPCR and EMSA assay, HUVECs were employed; and for Dual luciferase assay, 293T cells was employed.) *P < 0.05, ***P < 0.001 and ****P < 0.0001 (means±SDs; n=6) analyzed by one-way ANOVA with Tukey's multiple comparisons test (F, G, K and L) or Student’s t test (J) (ns, not significant). The exact P values are presented in the “Supplemental Statistical Table”.

Figure 5.. SUMOylation affects the formation of HEY1 dimer, thus altering its preference for the promoter-binding motif.

A, Detection of HEY1-WT/3KR homodimer formation in HUVECs. The relative dimer formation was normalized to that of the HEY1-WT dimer group. B, Detection of HEY1-WT/3KR-HEY2 heterodimer formation in HUVECs. The relative dimer formation was normalized to that of the HEY1-WT group. C, Detection of HEY1-WT/3KR-HES1 heterodimer formation in HUVECs. The relative dimer formation was normalized to that of the HEY1-WT group. D and E, mRNA and protein levels of HES1 in Ad-HEY1-WT/3KR-transduced HUVECs. The mRNA/protein level was normalized to that of β-ACTIN and the control group. F, Detection of the interaction between HEY1-WT/3KR and endogenous HES1 in HUVECs. The relative dimer formation was normalized to that of the HEY1-WT group. G, Interaction between HEY1-WT/3KR and N-box motifs decreased upon HES1 knockdown in HUVECs. The relative N-box binding was normalized to that of the si-Ctrl-HEY1-WT group. H, Interaction between HEY1 and N-box motifs decreased upon HES1 knockdown in HUVECs. The relative promoter enrichment was normalized to that of the input and of the control group. I, Schematic showing the preference of HEY1 for E-box or N-box motif. ****P < 0.0001 (means±SDs; n=6) analyzed by one-way ANOVA with Tukey's multiple comparisons test (A, D, E and H), two-way ANOVA with Sidak's multiple comparisons (G) or Student’s t test (B, C and F) (ns, not significant). The exact P values are presented in the “Supplemental Statistical Table”.

Figure 6.. SUMOylation facilitates HEY1 in the repression on of retinal angiogenesis.

A and B, Retinal angiogenesis in HEY1-WT/3KR^iEC-KI neonatal mice. Representative images and statistical analysis of vascular outgrowth on P7 in control, HEY1-WT^iEC-KI and HEY1-3KR^iEC-KI retinas stained with IB4 are shown. Scale bar: the upper panel, 1000 μm; the middle panel, 200 μm; and the lower panel, 100 μm. C and D, Proliferation of endothelial cells in P7 HEY1-WT/3KR EC-KI retinas. Representative images and statistical analysis of EdU+ IB4+ cells on P7 in control, HEY1-WT^iEC-KI and HEY1-3KR^iEC-KI retinas are shown. Scale bar, 50 μm. E and F, Phosphorylated VEGFR2 in P7 HEY1-WT/3KR EC-KI retinas. Representative images and statistical analysis of p-VEGFR2 area on P7 in control, HEY1-WT^iEC-KI and HEY1-3KR^iEC-KI retinas are shown. Scale bar, 50 μm. G and H, Retinal angiogenesis was repressed by SKLB610 (an inhibitor of p-VEGFR2) in HEY1-3KR but not HEY1-WT EC-KI neonatal mice. Representative images and statistical analysis of vascular outgrowth on P7 in HEY1-WT^iEC-KI and HEY1-3KR^iEC-KI retinas stained with IB4 are shown. Scale bar: the upper panel, 200 μm; and the lower panel, 50 μm. *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 (means±SDs; n=6) analyzed by one-way ANOVA with Tukey's multiple comparisons test (B, D and F) or two-way ANOVA with Sidak's multiple comparisons (H) (ns, not significant). P <0.0125 was considered statistically significant for multiple testing correction in B and P <0.025 was considered statistically significant for multiple testing correction in D and H based on the Bonferroni adjustment. The exact P values are presented in the “Supplemental Statistical Table”.

Figure 7.. SUMOylation of HEY1 in endothelial cells effectively compromises pathological angiogenesis.

A and B, Endothelial expression of HEY1-WT but not the HEY1-3KR mutant protects against aberrant neovascularization in the oxygen-induced retinopathy (OIR) model. Representative images and statistical analysis of vascular area on P17 in control, HEY1-WT^iEC-KI and HEY1-3KR^iEC-KI retinas stained with IB4 are shown. Scale bar: the upper and the middle panels, 1000 μm; the lower panel, 200 μm. C and D, Proliferation of endothelial cells and phosphorylation of VEGFR2 in P17 HEY1-WT/3KR retinas in the OIR model. Representative images and statistical analysis of p-VEGFR2+ CD31+ and EdU+ CD31+ cells area respectively, on P17 in control, HEY1-WT^iEC-KI and HEY1-3KR^iEC-KI retinas are shown. Scale bar: the upper panel, 200 μm; the lower panel, 50 μm. E and F, Endothelial expression of HEY1-WT but not the HEY1-3KR mutant is able to protect against preretinal neovascular tufts development in the OIR model. Representative images of H&E-stained sagittal sections of eyes from P17 mice and statistical analysis of preretinal neovascular tufts are shown. Scale bar, 500 μm. G, Representative images and statistical analysis of explant LLC tumors from control and HEY1-WT/3KR^iEC-KI mice. Scale bar, 5 mm. H, Representative images and statistical analysis of immunofluorescence staining with CD31, NG2, α-SMA, Collagen IV (Col4) antibody in explant LLC tumors from control and HEY1-WT/3KR^iEC-KI mice. Scale bar: 100 μm. *P < 0.05, ***P < 0.001 and ****P < 0.0001 (means±SDs; n=6 for B, D, F and H; n=8 for G) analyzed by one-way ANOVA with Tukey's multiple comparisons test (B, D, F and H) or two-way ANOVA with Tukey's multiple comparisons test (G) (ns, not significant). P <0.025 was considered statistically significant for multiple testing correction in B based on the Bonferroni adjustment. The exact P values are presented in “Supplemental Statistical Table”.

Figure 8.. The critical role of HEY1 SUMOylation in the regulation of angiogenesis.

A, Validation of downregulation of selected genes upon expression of HEY1-WT but not HEY1-3KR in Hey1^iEC-KO MLECs by quantitative PCR and immunoblotting. The mRNA/protein level was normalized to that of β-ACTIN and the control group. B, EMSA showing HEY1-WT/3KR binding to E-box and N-box motifs in MLECs. The relative E/N-box binding was normalized to that of the HEY1-WT^iEC-KI group. C, Schematic diagram illustrating the critical role of HEY1 SUMOylation in the regulation of angiogenesis. *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 (means±SDs; n=6) analyzed by one-way ANOVA with Tukey's multiple comparisons test (A) Student’s t test (B) (ns, not significant). The exact P values are presented in “Supplemental Statistical Table”.