Neuro-oncological Ventral Antigen 2 Regulates Splicing of Vascular Endothelial Growth Factor Receptor 1 and Is Required for Endothelial Function

Abstract

Pre-eclampsia (PE) affects 2-8% of pregnancies and is responsible for significant morbidity and mortality. The maternal clinical syndrome (defined by hypertension, proteinuria, and organ dysfunction) is the result of endothelial dysfunction. The endothelial response to increased levels of soluble FMS-like Tyrosine Kinase 1 (sFLT1) is thought to play a central role. sFLT1 is released from multiple tissues and binds VEGF with high affinity and antagonizes VEGF. Expression of soluble variants of sFLT1 is a result of alternative splicing; however, the mechanism is incompletely understood. We hypothesize that neuro-oncological ventral antigen 2 (NOVA2) contributes to this. NOVA2 was inhibited in human umbilical vein endothelial cells (HUVECs) and multiple cellular functions were assessed. NOVA2 and FLT1 expression in the placenta of PE, pregnancy-induced hypertension, and normotensive controls was measured by RT-qPCR. Loss of NOVA2 in HUVECs resulted in significantly increased levels of sFLT1, but did not affect expression of membrane-bound FLT1. NOVA2 protein was shown to directly interact with FLT1 mRNA. Loss of NOVA2 was also accompanied by impaired endothelial functions such as sprouting. We were able to restore sprouting capacity by exogenous VEGF. We did not observe statistically significant regulation of NOVA2 or sFLT1 in the placenta. However, we observed a negative correlation between sFLT1 and NOVA2 expression levels. In conclusion, NOVA2 was found to regulate FLT1 splicing in the endothelium. Loss of NOVA2 resulted in impaired endothelial function, at least partially dependent on VEGF. In PE patients, we observed a negative correlation between NOVA2 and sFLT1.

Keywords: Alternative Splicing; Endothelium; FLT1; Placenta; Pre-eclampsia.

Fig. 1

NOVA2 and FLT1 are enriched in ECs. A, C We obtained pelleted cells of different cell types in the cardiovascular system from Promocell and RNA sequencing was performed using different cell types from the cardiovascular system. A NOVA2 and C FLT1 are shown. B, D RNA was isolated from mouse hearts. Cardiomyocytes were separated from other cells by density centrifugation and used for RNA sequencing. B Nova2 and D Flt1 are shown. Groups were analyzed using an unpaired t-test. Three samples were included in each group. Data are presented as mean ± SEM. Significance was indicated as follows: *p < 0.05, **p < 0.01, ***p < 0.001, not significant (ns)

Fig. 2

Induction of sFLT1 but not FLT1 mRNA levels following NOVA2 silencing in HUVECs. A Schematic representation of FLT1 variants and binding site of RT-qPCR primers. B NOVA2 was silenced in HUVECs using siRNA. A non-targeting siRNA was transfected as a control. RNA was isolated 48 h after transfection and NOVA2 expression was measured by RT-qPCR. Expression is normalized to RPLP0. Groups were compared using a paired t-test. 3 independent experiments were performed. C NOVA2 protein levels were determined using Western blot. Cell lysates were collected 48 h after transfection. GAPDH was used as a loading control. Images were cropped and contrast was increased for clarity. 2 experiments were performed. D–G NOVA2 was silenced using siRNA. A non-targeting siRNA was transfected as a control. RNA was isolated 48 h after transfection and D total FLT1, E membrane FLT1, F sFLT1-i13, and G sFLT1-e15a expression was measured by RT-qPCR. Expression is normalized to RPLP0. Groups were compared using a paired t-test. 3–7 independent experiments were performed. H, I FLT1 binding to NOVA2 was analyzed in HUVECs by RT-qPCR following CLIP. Non-targeting IgG was used as a control. Enrichment was quantified relative to input. Groups were compared using a paired t-test. 6 independent experiments were performed. I NOVA2 pulldown was detected using Western blot. Image was cropped and contrast was increased for clarity. Of note, the IgG heavy chain detected at 50 kDa overlaps with the NOVA2 band. J Schematic representation of human FLT1. Exons are in blue; the transcript reads into intron 13 for sFLT1-i13 are in light blue. The highlighted YCAY sites are identified by RBPmap and SpliceAid2 within 500 nt from the exon–intron junction. For each potential binding site predicted by RBPmap, the Z-score > 2.52 and p < 5.57e–03. Data are presented as mean ± SEM. Significance was indicated as follows: *p < 0.05, **p < 0.01, ***p < 0.001, not significant (ns)

Fig. 3

Loss of NOVA2S results in impaired sprouting and mild barrier impairment. A EC spheroids were embedded in collagen gels 24 h after transfection and stimulated with VEGF. Fixation was done after 24-h VEGF stimulation. Cumulative sprout length was determined by measuring the distance from the base of the spheroid to the tip cell. Discontinuous sprouts were excluded. 10 experiments were performed; for each independent experiment, 7–12 spheroids were scored in each experiment. Scale bar indicates 250 μm. B–D HUVECs were seeded 24 h after transfection at a density of 100,000 cells/well in 8W10E ECIS plates. Impedance was measured continuously. By altering the frequency, overall barrier (B), cell–cell contact (C), and cell–matrix contact (D) can be distinguished. Groups were analyzed at t = 24 h using paired t-test. 7 independent experiments were performed. E Proliferation was measured by EdU incorporation between 24 and 48 h after transfection. The percentage of proliferating cells is shown. Groups were analyzed using a paired t-test. 3 experiments were performed. F HUVECs were seeded in a 96-well plate 45 h after transfection. Apoptosis was induced using 200-nm staurosporine. Caspase substrate was added and fluorescence was measured after 1 h. Fluorescence intensity for each sample was normalized to the total signal from the plate. Groups were analyzed using a paired one-way ANOVA. 4 experiments were performed. Data are presented as mean ± SEM. Significance was indicated as follows: *p < 0.05, **p < 0.01, ***p < 0.001, not significant (ns)

Fig. 4

Reduced sFLT1/mFLT1 ratio correlates with increased NOVA2 expression. A–D Frozen placenta tissue was obtained from 15 PE, 8 PIH patients, and 14 NT controls. RNA was isolated and A NOVA2, B mFLT1, C sFLT1-i13, and D sFLT1-e15a were measured by RT-qPCR. Expression is normalized to SDHA. Groups were compared using one-way ANOVA. E, F sFLT1 relative expression correlation with NOVA2 relative expression. Correlation was estimated using simple linear regression. Data are presented as mean ± SEM. Significance was indicated as follows: *p < 0.05, **p < 0.01, ***p < 0.001, not significant (ns)