Endothelin-1 Drives Epithelial-Mesenchymal Transition in Hypertensive Nephroangiosclerosis

Abstract

Background: Tubulointerstitial fibrosis, the final outcome of most kidney diseases, involves activation of epithelial mesenchymal transition (EMT). Endothelin-1 (ET-1) activates EMT in cancer cells, but it is not known whether it drives EMT in the kidney. We therefore tested the hypothesis that tubulointerstitial fibrosis involves EMT driven by ET-1.

Methods and results: Transgenic TG[mRen2]27 (TGRen2) rats developing fulminant angiotensin II-dependent hypertension with prominent cardiovascular and renal damage were submitted to drug treatments targeted to ET-1 and/or angiotensin II receptor or left untreated (controls). Expressional changes of E-cadherin and α-smooth muscle actin (αSMA) were examined as markers of renal EMT. In human kidney HK-2 proximal tubular cells expressing the ETB receptor subtype, the effects of ET-1 with or without ET-1 antagonists were also investigated. The occurrence of renal fibrosis was associated with EMT in control TGRen2 rats, as evidenced by decreased E-cadherin and increased αSMA expression. Irbesartan and the mixed ET-1 receptor antagonist bosentan prevented these changes in a blood pressure-independent fashion (P < 0.001 for both versus controls). In HK-2 cells ET-1 blunted E-cadherin expression, increased αSMA expression (both P < 0.01), collagen synthesis, and metalloproteinase activity (P < 0.005, all versus untreated cells). All changes were prevented by the selective ETB receptor antagonist BQ-788. Evidence for involvement of the Rho-kinase signaling pathway and dephosphorylation of Yes-associated protein in EMT was also found.

Conclusions: In angiotensin II-dependent hypertension, ET-1 acting via ETB receptors and the Rho-kinase and Yes-associated protein induces EMT and thereby renal fibrosis.

Keywords: endothelin‐1; epithelial to mesenchymal transition; fibrosis; hypertension; kidney.

Figure 1

Blood pressure in TGRen2 rats. Systolic blood pressure values in the TGRen2 rats at baseline (age 5 weeks) and after 4 weeks of placebo (controls; n = 8), or treatment with irbesartan (n = 5), bosentan (n = 5), BMS‐182874 (n = 4), and irbesartan plus BMS‐182874 (n = 5). At age 9 weeks control rats showed significantly increased blood pressure. The pressor increase was prevented only in the irbesartan group.

Figure 2

Markers of EMT in kidney of TGRen2 rats. A, Immunostaining for E‐cadherin (top panels), αSMA (middle panels), and S100A4 (bottom panels) of TGRen2 kidney sections after placebo treatment (a, f, k), irbesartan (b, g, l), bosentan (c, h, m), BMS 182874 (BMS) (d, i, n), and irbesartan on top of BMS (e, j, o). The specific signal pertaining to E‐cadherin is barely evident (empty arrows) in most tubules of placebo (a) and BMS‐treated (d) groups, where it is evident only in the vessel wall (asterisks). E‐cadherin can be clearly appreciated (full arrows) in the irbesartan‐ (b) and bosentan‐treated (c) groups. αSMA immunosignal is well recognizable in the proximal tubules of placebo (f) and BMS (i) groups but not in irbesartan‐ (g), bosentan‐ (h), and irbesartan+BMS‐ (j) treated groups. Note that tubules are enlarged in large part of the sections of placebo TGRen2 kidneys (f). The immunoreaction for S100A4 was evident in the tubules of placebo (k) and BMS (n) groups, but not in irbesartan‐ (l) and bosentan‐treated (m) groups. As expected, the immunosignal was clearly appreciated in the glomeruli. Inset, The omission of the primary antibody confirmed the specificity of the reaction. B, Quantitative analysis of immunostaining for the epithelial marker E‐cadherin and the mesenchymal markers αSMA and S100A4 in the kidney of TGRen2 rats after exposure to in vivo treatment with irbesartan, bosentan, BMS, and irbesartan on top of BMS. E‐cadherin (left panel) increased, while αSMA (middle panel) and S100A4 (right panel) decreased, after irbesartan and bosentan treatment, as compared to placebo. Irbesartan on top of BMS, but not BMS alone, prevented αSMA and S100A4 increases. Specific immunoreactivity was estimated as the percentage of total surface area. Only significant comparisons were reported. The number of animals is shown in Figure 1.

Figure 3

Coexpression of the epithelial and mesenchymal markers in the kidney of TGRen2 rats. Double immunofluorescence for E‐cadherin (green) and αSMA (red), or E‐cadherin (green) and S100A4 (red) in placebo TGRen2 kidney sections that include proximal tubules. Double immunofluorescence indicates coexpression of the epithelial and mesenchymal markers, thereby suggesting EMT. Tubules showed coexpression of αSMA and E‐cadherin (A and B; yellow, arrows) in few cells, thereby suggesting, as expected, that EMT occurred in only a small number of cells. However, some tubules showed only signal for αSMA, suggesting replacement of epithelial cells with myofibroblastoid cells throughout the tubule circumference, indicating overt fibrosis (C). Coexpression of S100A4 with E‐cadherin was rarely detected (D and E; arrows). The specificity of the immunoreaction for each marker was confirmed by the lack of signal after omission of primary antibody (inset). Coexpression of E cadherin and αSMA (F–J), or E cadherin and S100A4 (K–O) was restricted to vessels (asterisk) in irbesartan‐ and bosentan‐treated rats (G, L, M), thereby suggesting that EMT in the nephrons was fully prevented by either AT1 antagonism or nonselective ET_A/ET_B antagonism. In contrast, coexpression of epithelial and mesenchymal markers (yellow) was evident in some epithelial tubular cells after BMS (I; arrow), thereby suggesting that ET_A antagonism failed to prevent EMT. Irbesartan on top of BMS prevented coexpression of epithelial and mesenchymal markers, indicating the beneficial effect of AT1 antagonism. Yellow or orange‐reddish staining indicates double labeling. ×20 magnification.

Figure 4

Time course of gene expression of E‐cadherin and αSMA in HK‐2 cells after exposure to ET‐1. A, E‐cadherin gradually decreased until it reached a nadir at 48 to 72 hours after the ET‐1 challenge, and it returned to pretreatment levels after 96 hours. B, αSMA increased at 24 hours and at 96 hours, but its changes followed a less consistent course. It was back to pretreatment levels by day 10. Values are means of 4 independent experiments, each done at least in duplicate. P < 0.05, ANOVA with Bonferroni post‐hoc test for multiple comparisons.

Figure 5

Time course of the changes of E‐cadherin (A), αSMA (B), and vimentin (C) protein expression in HK‐2 cells exposed to ET‐1 or TGFβ. Both 10⁻⁷ mol/L ET‐1 and 2 nmol/L TGFβ, a well‐known inducer of EMT used as a positive control, caused a progressive decrease of E‐cadherin and an increase of αSMA and vimentin protein expression. Values are expressed as fold changes of the values recorded in untreated cells at time 0, set to 1 (0: n = 5; 7 days: n = 6; 10 days: n = 5, n: independent experiments, each at least in duplicate). Percentage changes of protein expression of E‐cadherin (D), αSMA (E), and vimentin (F) over control untreated cells, set to 0, after exposure of HK‐2 cells to ET‐1 for 10 days, or to macitentan and the selective ET_B receptor antagonist BQ‐788 on top of ET‐1. Asterisks indicate comparisons between ET‐1 and control untreated cells (*P < 0.05; **P < 0.01). Values are means of 3 independent experiments, each done at least in duplicate.

Figure 6

Disruption of cell junctions (panel A), onset of αSMA (panel B), expression of ET receptor subtypes (panel C), and changes of collagens after exposure of HK‐2 cells to ET‐1 (panel D). A, Confocal analysis of HK‐2 cells, which revealed integrity of the tight junctions immunolabeled in green with antibody against ZO‐1. After treatment with ET‐1, the junctions appeared to be partially disrupted, and αSMA (red) was evident in some cells. B, Onset of αSMA (red) was prevented by macitentan pretreatment. ZO‐1 and the nuclei are labeled in green and blue, respectively. C, ET_B was expressed more markedly than ET_A in HK‐2 cells, with a ratio ET_A/ET_B equal to 4.5. The omission of primary antibody confirmed the specificity of the reaction (inset). D, Gene expression of collagen 1α (COL1A), 2α (COL2A), 3α (COL3A), and 4α (COL4A) in HK‐2 cells at 24 hours, 48 hours, and 7 days after exposure to ET‐1 (4 independent experiments, each in duplicate). The value measured at time 0 for each gene set at 1 was used as control. *P = 0.05; ***P = 0.005, vs time 0.

Figure 7

Gene expression of metalloproteinases (MMP) (panel A), immunofluorescence of MMP9 (panel B), protein expression of MMP9 (panel C), and cell migration after exposure of HK‐2 cells to ET‐1 (panel D). A, MMP‐2 transiently increased after 12 hours of exposure to ET‐1 (4 independent experiments, each in duplicate). No significant differences between times were found for MMP‐9 gene expression. B, MMP9 (red) in HK‐2 cells after incubation was evident after treatment with ET‐1 (top) and blunted after pretreatment with macitentan (bottom). The omission of primary antibody confirmed the specificity of the reaction (inset). C, MMP9 increased after treatment with ET‐1 for 7 days, and macitentan was effective in preventing its expression (3 independent experiments, each in duplicate). D, Because gene and/or protein expression cannot parallel enzymatic activity of MMP, we measured MMP9 protein activity in HK‐2 cells exposed to ET‐1 or macitentan or BQ‐788 on top of ET‐1 for 24 hours. Untreated cells were used for comparison (3 independent experiments, each in duplicate). TGFβ was used as positive control. E, Representative curves showing cell migration, as assessed with cell–electrode impedance detection‐based technology. ET‐1 (turquoise) was able to induce migration, similar to that induced by the well‐known EMT inducer TGFβ (green). Untreated cells (purple) were used as negative controls. ET‐1–induced cell migration was abolished by pretreatment with macitentan (red) and BQ‐788 (blue). A.U., arbitrary units. F, Cell index values recorded 12 hours after the treatment. Mean values of 3 independent experiments, each in duplicate. *P = 0.01.

Figure 8

Changes of phosphorylation of MYPT‐1 after exposure of HK‐2 cells to ET‐1. A, Time course of MYPT‐1 phosphorylation showed that the ratio of phosphorylated MYPT‐1 to MYPT‐1 increased, as expected, early, ie, after 5 and 30 minutes of exposure of HK‐2 cells to ET‐1. Macitentan effectively prevented such an effect. B, BQ‐788 also prevented the increase in the phosphorylated MYPT‐1/MYPT‐1 ratio induced by ET‐1. Mean values of 3 independent experiments, each in duplicate.

Figure 9

YAP activation and intracellular localization of YAP after exposure of HK‐2 cells to ET‐1. A, Time course of YAP activation, expressed as the ratio of YAP to phosphorylated YAP, showed YAP activation 30 minutes after ET‐1 exposure. Graph bars represent mean values ± SEM. B, Immunofluorescence for YAP (green) in HK‐2 cells after exposure to ET‐1, in the presence or absence of either macitentan or BQ‐788. In untreated HK‐2 cells the immunosignal was clearly evident in both nucleus and cytoplasm. ET‐1 induced blunting of the YAP signal in the cytoplasm and its increase in the nucleus, indicating dephosphorylation of YAP and translocation to the nucleus, where it can trigger transcription of genes involved in EMT. Pretreatment with either macitentan or BQ‐788 prevented the ET‐1–induced increase of YAP immunosignal in the nucleus. Nuclei were labeled with DAPI (blue); the omission of primary antibody confirmed the specificity of the reaction (insets). C, Intracellular YAP localization. Quantitative analysis confirmed that ET‐1 favors translocation of YAP to the nucleus (vs untreated cells) and that macitentan and BQ‐788 prevented such translocation. For all experiments mean values of 3 independent experiments, each in duplicate, are reported.

Figure 10

ET‐1 drives EMT and consequently fibrosis in the kidney. The cartoon depicts the epithelial‐to‐mesenchymal phenotypic shift induced by ET‐1 in the renal tubules. Tubular cells are tightly linked to adjacent cells by junctions (visualized as red dashes) under physiologic conditions. On binding of ET‐1, the ET_B receptor subtype triggers a cell phenotype switch by which cells start expressing mesenchymal markers such as αSMA and vimentin and gradually lose their epithelial markers, such as E‐cadherin, which is the major component of adherent junctions. The phenotypic switch is represented in this cartoon as a change of color from bluish‐violet to greenish. Hence, coexistence of blue and greenish denotes an intermediate phenotype, whereas greenish indicates a complete transition to the mesenchymal phenotype. The cells undergoing this switch also exhibit a loss of cell junctions alongside activation of MMP‐9, which promotes degradation of the basal membrane. These concomitant processes allow the modified cells to move toward the interstitium, where they start producing extracellular matrix proteins, including collagen (red and blue lines) that, by accumulating in the interstitium between tubules, leads to tubulointerstitial fibrosis. The magnified tubular cell illustrates in greater detail the intracellular pathways activated by ET‐1. Phosphorylation of MYPT, a member of the Rho/Rock pathway that leads to cytoskeleton contraction and cell motility, is activated, whereas YAP phosphorylation is prevented, thereby allowing entry of YAP into the nucleus and finally leading to transcription of target genes that promote cell migration. Dashed lines indicate yet not proven effects.