Potential Role of Circulating Endoglin in Hypertension via the Upregulated Expression of BMP4

Abstract

Endoglin is a membrane glycoprotein primarily expressed by the vascular endothelium and involved in cardiovascular diseases. Upon the proteolytic processing of the membrane-bound protein, a circulating form of endoglin (soluble endoglin, sEng) can be released, and high levels of sEng have been observed in several endothelial-related pathological conditions, where it appears to contribute to endothelial dysfunction. Preeclampsia is a multisystem disorder of high prevalence in pregnant women characterized by the onset of high blood pressure and associated with increased levels of sEng. Although a pathogenic role for sEng involving hypertension has been reported in several animal models of preeclampsia, the exact molecular mechanisms implicated remain to be identified. To search for sEng-induced mediators of hypertension, we analyzed the protein secretome of human endothelial cells in the presence of sEng. We found that sEng induces the expression of BMP4 in endothelial cells, as evidenced by their proteomic signature, gene transcript levels, and BMP4 promoter activity. A mouse model of preeclampsia with high sEng plasma levels (sEng⁺) showed increased transcript levels of BMP4 in lungs, stomach, and duodenum, and increased circulating levels of BMP4, compared to those of control animals. In addition, after crossing female wild type with male sEng⁺ mice, hypertension appeared 18 days after mating, coinciding with the appearance of high plasma levels of BMP4. Also, serum levels of sEng and BMP4 were positively correlated in pregnant women with and without preeclampsia. Interestingly, sEng-induced arterial pressure elevation in sEng⁺ mice was abolished in the presence of the BMP4 inhibitor noggin, suggesting that BMP4 is a downstream mediator of sEng. These results provide a better understanding on the role of sEng in the physiopathology of preeclampsia and other cardiovascular diseases, where sEng levels are increased.

Keywords: BMP4; HHT; TGF-β; endoglin; endothelial cells; hypertension; preeclampsia.

Figure 1

iTRAQ identification of differentially secreted proteins by human umbilical vein endothelial cells (HUVECs) in the presence of sEng. (A) Volcano type graph where secreted proteins from sEng-treated HUVECs (HUVECs-sEng) that change significantly with respect to the untreated control (HUVECs-Control) are indicated. The horizontal axis represents fold changes in induction of the ratio HUVECs-sEng/HUVECs-Control, considering negligible differences those values closer to 0. The vertical axis represents the -Log10 (p-value), and the continuous horizontal line plotted at the value of 2.3, is equivalent to a p-value of 0.005. The central gray point cloud represents quantified but not statistically significant proteins. Inhibited or over-expressed proteins with statistically significant differences after treatment with sEng are represented in blue (upper left quadrant) or red (upper right quadrant), respectively. (B) Table showing the list of proteins represented in the graph in which statistically significant differences were found, including the corresponding access number in the UniProt database (http://www.uniprot.org/uniprot/).

Figure 2

BMP4 expression upon in vitro treatment of cells with sEng. (A,B) HUVECs were treated with sEng at the indicated concentrations. Protein levels of BMP4 secreted into the medium were determined by ELISA and normalized to control (A). Transcript levels of BMP4 were measured by qRT-PCR and values normalized to the control condition (B). (C) sEng transactivates the BMP4 promoter. HEK293T cells were transiently transfected with the pEZX-PG04.1 vector encoding the Gaussia luciferase driven by the human BMP4 promoter. Cell transfections were performed in the presence of increasing concentrations of sEng and after forty-eight hours, the luciferase activity of cell lysates was measured by luminometry. Unless otherwise indicated, p values are referred to the control condition. (* p < 0.05; ** p < 0.01).

Figure 3

Correlation between sEng and BMP4 levels in sera from pregnant women. Sera from pregnant women without (control; n = 19) or with preeclampsia (n = 16) were analyzed by ELISA to determine protein levels of sEng (A; *** p = 0.001), BMP4 (B; p = 0.13) and sFlt1. In panel B, the presence of high-outliers (*^,°) is indicated. Correlation analysis between sEng and BMP4 levels (C; r = 0.35, p = 0.03), and between sEng and sFlt1 levels (D; r = 0.81, p < 0.0001) are shown.

Figure 4

Expression of BMP4 in sEng⁺ transgenic mice. (A,B) Plasma levels of BMP4 and sEng. (A) Phenotyping of mice used in the study. Plasma levels of sEng present in WT (n = 35) and sEng⁺ (n = 35) animals were determined by ELISA. (B) Plasma levels of BMP4 in WT (n = 6) and sEng⁺ (n = 5) mice, as measured by ELISA. The mean and the standard error of the mean are represented. (*** p < 0.001). (C) Immunohistochemical staining of BMP4 in mouse tissues. Lung, duodenum and stomach sections from WT or sEng⁺ mice were stained with anti-BMP4 (brown color) and counterstained hematoxylin (blue color), as described in Materials and Methods. The images were taken at ×20 magnification. Scale bars, 50 µm. (D) Gene expression levels of BMP4 in mouse tissues. Quantification by qRT-PCR of BMP4 mRNA expression in lung (n = 7 animals for each genotype), duodenum (n = 7 WT mice and n = 8 sEng⁺ mice), stomach (n = 5 WT mice and n = 7 sEng⁺ mice) from WT or sEng⁺ mice. Results were normalized, using 18R ribosomal RNA expression as internal control. The mean and the standard error are represented. (*** p < 0.001).

Figure 5

Effect of noggin administration on systolic blood pressure. Pumps loaded with either vehicle or noggin were implanted in WT or sEng⁺ mice and four days later, arterial pressure (A) as well as plasma BMP4 (B) and sEng (C) levels were measured. n = 4 in each noggin-treated group; n = 3 in each vehicle-treated group. (* p < 0.05; ** p < 0.01; *** p < 0.001).

Figure 6

Analysis of plasma sEng and BMP4 levels and blood pressure in a model of preeclampsia. Wild type (WT) female mice were crossed with either sEng⁺ or WT males. The resulting WT pregnant females bearing sEng⁺ fetuses [fWT(sEng⁺)] or WT fetuses [fWT(WT)] were analyzed for sEng (A) and BMP4 (C) plasma levels, as well as SBP (B) at days 13 and 18 after pregnancy. (A,C) n = 9 mice for each group. (B) n = 14 mice for each group. (* p < 0.05; *** p < 0.001).

Figure 7

Hypothetical model for BMP4 in sEng-dependent effect in hypertension. Upon inflammation, MMP12 and/or MMP14 trigger the release of sEng in endothelial cells by proteolytically cleaving membrane-bound endoglin. In turn, sEng stimulates the expression of BMP4 in different organs, including lung, stomach and duodenum. Next, circulating BMP4 contributes to endothelial dysfunction, leading to hypertension.