ABSTRACT

Endothelial-to-mesenchymal transition (EndMT) is characterized by the loss of endothelial cell markers and functions, and coincides with de novo expression of mesenchymal markers. EndMT is induced by TGFβ1 and changes endothelial microRNA expression. We found that miR-20a is decreased during EndMT, and that ectopic expression of miR-20a inhibits EndMT induction. TGFβ1 induces cellular hypertrophy in human umbilical vein endothelial cells and abrogates VE-cadherin expression, reduces endothelial sprouting capacity and induces the expression of the mesenchymal marker SM22α (also known as TAGLN). We identified ALK5 (also known as TGFBR1), TGFBR2 and SARA (also known as ZFYVE9) as direct miR-20a targets. Expression of miR-20a mimics abrogate the endothelial responsiveness to TGFβ1, by decreasing ALK5, TGFBR2 and SARA, and inhibit EndMT, as indicated by the maintenance of VE-cadherin expression, the ability of the cells to sprout and the absence of SM22α expression. FGF2 increases miR-20a expression and inhibits EndMT in TGFβ1-stimulated endothelial cells. In summary, FGF2 controls endothelial TGFβ1 signaling by regulating ALK5, TGFBR2 and SARA expression through miR-20a. Loss of FGF2 signaling combined with a TGFβ1 challenge reduces miR-20a levels and increases endothelial responsiveness to TGFβ1 through elevated receptor complex levels and activation of Smad2 and Smad3, which culminates in EndMT.

INTRODUCTION

Endothelial–mesenchymal transition (EndMT) is a process where endothelial cells lose their endothelial characteristics and start to differentiate into mesenchymal cells (Krenning et al., 2008, 2010; Moonen et al., 2010; Maleszewska et al., 2013). EndMT is evidenced by the phenotypic and functional alteration of endothelial cells, which results in the repression of endothelial cell markers [e.g. VE-cadherin and CD31 (also known as PECAM1)] and functions (e.g. sprouting ability and thrombolytic properties) and the de novo expression mesenchymal proteins [e.g. SM22α (also known as TAGLN) and αSMA (also known as ACTA2)] and acquirement of mesenchymal cell functions (i.e. contractile behavior and collagen production) (Krenning et al., 2008; Moonen et al., 2010; Maleszewska et al., 2013).

Phenotypic transitions are pivotal to embryonic processes such as tissue differentiation and morphogenesis, and, of these, EndMT is essential in the formation of the cardiac valves (Krenning et al.,2010). However, postnatally, phenotypic transitions are linked to pathological processes, such as tumor metastasis (Potenta et al., 2008) and the occurrence of fibroproliferative diseases. EndMT has been shown to contribute to the formation of neointimal lesions (Cooley et al., 2014; Moonen et al.,2015), vascular calcifications (Yao et al., 2013), and fibrosis of the lungs (Hashimoto et al., 2010), heart (Zeisberg et al., 2007) and kidneys (Zeisberg et al., 2008), as well as heterotypic ossifications (Medici and Olsen, 2012).

In recent years, it has become clear that signaling through the transforming growth factor-β (TGFβ) superfamily is key in EndMT (Goumans et al., 2008). TGFβ binds to the heteromeric receptor complex formed by the activin-like kinase 5 (ALK5, also known as TGFβ type I receptor, TGFBR1) and the TGFβ type II receptor (TGFBR2). Phosphorylation of ALK5 by the kinase TGFBR2, activates its catalytic kinase domain, allowing the activation of the receptor-regulated Smads (i.e. Smad2 and Smad3) (Goumans et al., 2008; Yoshimatsu and Watabe, 2011).

Smad anchor for receptor activation (SARA; also known as ZFYVE9), functions to recruit Smad2 and Smad3 to the TGFβ receptor complex, by controlling their subcellular localization and interacting with TGFBR1 (Tsukazaki et al., 1998). Activation of Smad2 and Smad3 causes their dissociation from SARA, a concomitant complexation with Smad4 and finally nuclear translocation and transcriptional regulation of target genes such as that encoding SM22α (Qiu et al., 2003). Although a clear role for TGFβ has been established in the induction of EndMT, less is known about the signaling mechanisms that modulate or inhibit the induction of EndMT.

Non-coding microRNAs (miRs) are post-transcriptional repressors of gene function and are often dysregulated in pathological processes (reviewed in Quiat and Olson, 2013). MiRs have diverse functions, including the regulation of cellular differentiation, proliferation and apoptosis. We have previously performed microRNA expression analysis by microarray to identify miRs that are deregulated during EndMT, and found that miR-20a expression is abolished during EndMT (our unpublished data). We also previously found that exogenous expression of miR-20a in endothelial cells potently inhibited EndMT induction.

MicroRNA-20a is encoded on chromosome 13 and part of the miR-17-92 cluster. Here, we show that fibroblast growth factor-2 (FGF2) induces the endothelial expression of miR-20a, which targets ALK5, TGFBR2 and SARA to inhibit canonical TGFβ signaling. Ectopic expression of miR-20a by endothelial cells blocks Smad2 and Smad3 activation and protects the endothelium from EndMT.

RESULTS

TGFβ1 induces endothelial–mesenchymal transition in an ALK5-dependent manner

Endothelial cells stimulated with TGFβ1 drastically changed their morphology and showed signs of cellular hypertrophy and reduced cell numbers (Fig. 1A,B). These morphological changes were accompanied by phenotypic alterations. TGFβ1-stimulated endothelial cells had a pronounced decrease in VE-cadherin expression (6.8-fold decrease, P<0.001, Fig. 1D–G) and increased their expression of the mesenchymal marker SM22α (7.9-fold increase, P<0.001, Fig. 1H–K), suggestive of EndMT. Endothelial cells that underwent EndMT lost the endothelial sprouting behavior (>60-fold reduction, P<0.001, Fig. 1L–O), corroborating functional alterations previously described during EndMT (Krenning et al., 2008; Moonen et al., 2010). Blockade of canonical TGFβ signaling with the small-molecule ALK5 inhibitor SB431542 inhibited these morphological, phenotypical and functional alterations (Fig. 1).

Fig. 1.

TGFβ1 induces EndMT in a manner that is dependent on ALK5, Smad2 and Smad3. (A–C) Light microscopy images of endothelial cells. (A) Untreated cells. Treatment of endothelial cells with TGFβ1 (B) induces hypertrophy, which is counteracted by treatment with the ALK5 inhibitor SB431542 (5 µM, C). Immunofluorescence analysis of VE-cadherin (D–F) and SM22α (H–J). (D,H) Untreated cells. TGFβ1 treatment of endothelial cells decreases VE-cadherin expression (E) and induces SM22α expression (I). The addition of SB431542 to the TGFβ1-treated cells maintains VE-cadherin expression (F) and inhibits SM22α expression (J). (G,K) Quantification of immunofluorescence analysis (n=5 per group). Matrigel sprouting of endothelial cells (L–N) is inhibited by TGFβ1 treatment (M) and restored by the addition of SB431542 (N). (L) Untreated cells. (O) Quantification of Matrigel sprouting ability of endothelial cells (n=5 per group). (P,Q) Representative immunoblots of components of endothelial TGFβ signaling (P) and their quantification (n=5 per group, Q). Graphical results are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001 (one-way ANOVA followed by Bonferroni post-tests).

Fig. 1.

TGFβ1 induces EndMT in a manner that is dependent on ALK5, Smad2 and Smad3. (A–C) Light microscopy images of endothelial cells. (A) Untreated cells. Treatment of endothelial cells with TGFβ1 (B) induces hypertrophy, which is counteracted by treatment with the ALK5 inhibitor SB431542 (5 µM, C). Immunofluorescence analysis of VE-cadherin (D–F) and SM22α (H–J). (D,H) Untreated cells. TGFβ1 treatment of endothelial cells decreases VE-cadherin expression (E) and induces SM22α expression (I). The addition of SB431542 to the TGFβ1-treated cells maintains VE-cadherin expression (F) and inhibits SM22α expression (J). (G,K) Quantification of immunofluorescence analysis (n=5 per group). Matrigel sprouting of endothelial cells (L–N) is inhibited by TGFβ1 treatment (M) and restored by the addition of SB431542 (N). (L) Untreated cells. (O) Quantification of Matrigel sprouting ability of endothelial cells (n=5 per group). (P,Q) Representative immunoblots of components of endothelial TGFβ signaling (P) and their quantification (n=5 per group, Q). Graphical results are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001 (one-way ANOVA followed by Bonferroni post-tests).

We analyzed the expression of proteins involved in canonical TGFβ signaling (Fig. 1P,Q). Stimulation of endothelial cells with TGFβ1 increased the expression of proteins in the TGFβ receptor complex, that is, ALK5 (2.6-fold, P<0.01), TGFBR2 (2.7-fold, P<0.01) and SARA (2.1-fold, P<0.001) and increased activation (i.e. phosphorylation) of the Smad2 and Smad3 transcription factors [phosphorylated (p)Smad2: 2.9-fold, P<0.01; pSmad3: 1.8-fold, P<0.01]. Activation of TGFβ signaling was reduced to baseline levels by the addition of 5 µM of the small-molecule ALK5 inhibitor SB431542 (Fig. 1P,Q).

These data indicate that active endothelial TGFβ signaling through ALK5 and the Smad2 and Smad3 transcription factors is associated with EndMT and confirms previous data reported by us and others (Kokudo et al., 2008; Krenning et al., 2008; Moonen et al., 2010; Medici et al., 2010, 2011; Yoshida et al., 2012).

miR-20a targets TGFβ signaling at the receptor level

We have previously observed that miR-20a is a potent inhibitor of EndMT and questioned its mechanism of inhibition. Online bioinformatics tools (i.e. the miRanda algorithm; Grimson et al., 2007; Betel et al., 2008) indicate that all components of the canonical TGFβ pathway are putative targets of miR-20a (Fig. 2A) because they contain one or more conserved binding sites for miR-20a. However, putative miR-20a targets in the TGFβ pathway vary highly in their miRSVR score, which indicates the likelihood of actual downregulation (Fig. 2A) (Betel et al., 2008, 2010).

Fig. 2.

miR-20a targets TGFβ­ signaling at the receptor level. (A) In silico analysis (microrna.org) of the TGFβ receptors and downstream mediators identifies multiple putative miR-20a-binding sites in the 3′UTR of these genes (red). Putative targeting efficiency is summarized in the miRSVR Score as previously described (Betel et al., 2008, 2010). (B) Luciferase reporter assays (n=6 per group) for putative miR-20a target genes identifies ALK5, TGFBR2 and SARA as genuine miR-20a targets, as co-transfection of these reporter constructs with miR-20a mimics in COS-7 cells reduces luciferase activity. Results are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001; ns, not significant (one-way ANOVA followed by Bonferroni post-tests).

Fig. 2.

miR-20a targets TGFβ­ signaling at the receptor level. (A) In silico analysis (microrna.org) of the TGFβ receptors and downstream mediators identifies multiple putative miR-20a-binding sites in the 3′UTR of these genes (red). Putative targeting efficiency is summarized in the miRSVR Score as previously described (Betel et al., 2008, 2010). (B) Luciferase reporter assays (n=6 per group) for putative miR-20a target genes identifies ALK5, TGFBR2 and SARA as genuine miR-20a targets, as co-transfection of these reporter constructs with miR-20a mimics in COS-7 cells reduces luciferase activity. Results are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001; ns, not significant (one-way ANOVA followed by Bonferroni post-tests).

Reporter assays, wherein the 3′UTR regions of miR-20a gene targets are coupled to a luciferase-encoding gene, revealed that the TGFβ signaling members ALK5, TGFBR2 and SARA are genuine miR-20a target genes (Fig. 2B) as co-transfection of these reporter plasmids with miR-20a mimics reduced luciferase activity by 1.4-fold (P<0.05), 1.7-fold (P<0.01) and 1.6-fold (P<0.001), respectively. In contrast, co-transfection of reporter plasmids with a scrambled miR-20a sequence did not alter luciferase activity (Fig. 2B).

Smad3 and Smad4 are putative gene targets of miR-20a (Fig. 2A); however, co-transfection of their respective reporter constructs with miR-20a mimics in COS7 cells, did not affect luciferase activity (Fig. 2B) suggesting that Smad3 and Smad4 are not genuine targets of miR-20a. Smad2 reporters had reduced luciferase activity upon co-transfection of the 3′UTR reporter construct with the miR-20a mimic, however only to a minimal extend (1.1-fold; P<0.05).

Combined, these data indicate that miR-20a affects canonical TGFβ signaling in endothelial cells at the level of the receptors, which might reduce downstream signaling.

miR-20a gain-of-function inhibits canonical TGFβ signaling

To validate that miR-20a affects endogenous TGFβ signaling, and thus EndMT, in endothelial cells, we transfected endothelial cells with miR-20a mimics (∼4-fold increase, P<0.001) or expressed scrambled control sequences (Fig. 3A). Endothelial cells treated with scrambled sequences and TGFβ1 underwent EndMT, as indicated by changes in cell morphology (Fig. 3C), waning of VE-cadherin expression (4.9-fold decrease, P<0.001; Fig. 3G,J) and the induction of SM22α (6.8-fold increase, P<0.001; Fig. 3L,O). Gain-of-function of miR-20a in endothelial cells did not affect the endothelial morphology or phenotype (Fig. 3D,H,M). Notably, miR-20a mimics did affect the endothelial responsiveness to the TGFβ challenge. Cells with ectopic expression of miR-20a were irresponsive to TGFβ1 and did not undergo EndMT, as indicated by the absence of hypertrophy (Fig. 3E), maintenance of VE-cadherin expression (Fig. 3I) and absence of SM22α expression (Fig. 3N). Moreover, miR-20a gain-of-function in TGFβ1-treated endothelial cells partially rescued the endothelial sprouting ability (2.4-fold increase, P<0.001) compared to scrambled controls (Fig. 3Q,S,T). TGFβ-induced EndMT was associated with increased mRNA expression of the mesenchymal transcription factors Snai1 (2.9-fold increase, P<0.001), Snai2 (4.5-fold increase, P<0.001) and Twist1 (4.4-fold increase, P<0.001), which were absent in the miR-20a-treated cells (Fig. 3U).

Fig. 3.

miR-20a gain-of-function inhibits TGFβ1, ALK5 and Smad2 and Smad3 signaling. (A) HUVECs were transfected with miR-20a mimics or scrambled (Scr) control sequences and challenged with TGFβ1. TGFβ1 readily decreased miR-20a levels in control cells, but not in cells transfected with miR-20a mimics. (B–E) Light microscopy images of endothelial cells. (B) Untreated cells. Treatment of endothelial cells with TGFβ1 (C) induces hypertrophy, which is counteracted by miR-20a mimics (E). Treatment of endothelial cells with miR-20a mimics did not alter cellular morphology (D). Immunofluorescence analysis of VE-cadherin (F–I) and SM22α (K–N). (F,K) Untreated cells. TGFβ1 treatment of endothelial cells decreases VE-cadherin expression (G) and induces SM22α expression (L). Transfection of HUVECs with miR-20a mimics maintains VE-cadherin expression (I) and inhibits SM22α expression (N) in TGFβ1-treated endothelial cells. Treatment of endothelial cells with miR-20a mimics alone did not alter VE-cadherin (H) expression nor SM22α expression (M). (J,O) Quantification of immunofluorescence analysis (n=5 per group). (P–T) Matrigel sprouting of endothelial cells (P–S) is inhibited by TGFβ1 treatment (Q) and restored by the addition of miR-20a mimics in endothelial cells (S). (P) Untreated cells. (T) Quantification of Matrigel sprouting ability of endothelial cells (n=5 per group). (U) Gene expression data on mesenchymal transcription factors, Snai1, Snai2 and Twist1. Expression is induced by TGFβ and inhibited by the addition of miR-20a. (V,W) Representative immunoblots of components of endothelial TGFβ signaling (V) and their quantification (n=5 per group, W). TGFβ1 increases expression of the TGFβ receptor complex and activates Smad2 and Smad3. These expression and activity changes are inhibited by the addition of miR-20a mimics to endothelial cells. Graphical results are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001 (one-way ANOVA followed by Bonferroni post-tests).

Fig. 3.

miR-20a gain-of-function inhibits TGFβ1, ALK5 and Smad2 and Smad3 signaling. (A) HUVECs were transfected with miR-20a mimics or scrambled (Scr) control sequences and challenged with TGFβ1. TGFβ1 readily decreased miR-20a levels in control cells, but not in cells transfected with miR-20a mimics. (B–E) Light microscopy images of endothelial cells. (B) Untreated cells. Treatment of endothelial cells with TGFβ1 (C) induces hypertrophy, which is counteracted by miR-20a mimics (E). Treatment of endothelial cells with miR-20a mimics did not alter cellular morphology (D). Immunofluorescence analysis of VE-cadherin (F–I) and SM22α (K–N). (F,K) Untreated cells. TGFβ1 treatment of endothelial cells decreases VE-cadherin expression (G) and induces SM22α expression (L). Transfection of HUVECs with miR-20a mimics maintains VE-cadherin expression (I) and inhibits SM22α expression (N) in TGFβ1-treated endothelial cells. Treatment of endothelial cells with miR-20a mimics alone did not alter VE-cadherin (H) expression nor SM22α expression (M). (J,O) Quantification of immunofluorescence analysis (n=5 per group). (P–T) Matrigel sprouting of endothelial cells (P–S) is inhibited by TGFβ1 treatment (Q) and restored by the addition of miR-20a mimics in endothelial cells (S). (P) Untreated cells. (T) Quantification of Matrigel sprouting ability of endothelial cells (n=5 per group). (U) Gene expression data on mesenchymal transcription factors, Snai1, Snai2 and Twist1. Expression is induced by TGFβ and inhibited by the addition of miR-20a. (V,W) Representative immunoblots of components of endothelial TGFβ signaling (V) and their quantification (n=5 per group, W). TGFβ1 increases expression of the TGFβ receptor complex and activates Smad2 and Smad3. These expression and activity changes are inhibited by the addition of miR-20a mimics to endothelial cells. Graphical results are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001 (one-way ANOVA followed by Bonferroni post-tests).

Gain-of-function of miR-20a in endothelial cells inhibited the TGFβ1-induced increase in protein expression of ALK5 (2.3-fold reduction, P<0.01), TGFBR2 (1.6-fold reduction, P<0.01) and SARA (2.8-fold reduction, P<0.01), which remained equal to the expression levels in untreated endothelial cells. As a consequence of decreased receptor availability in TGFβ1-treated endothelial cells, TGFβ1 was no longer able to activate Smad2 (2.4-fold reduction, P<0.05) and Smad3 (2.8-fold reduction, P<0.001) in miR-20a efficient cells (Fig. 3V,W). Interestingly, expression levels of ALK5, TGFBR2 and SARA showed a tendency to decrease (P<0.10) in endothelial cells treated with miR-20a mimics, but did not alter the levels of Smad2 and Smad3 activation (Fig. 3V,W).

Combined, these data show that miR-20a targets TGFβ signaling at the receptor level and the decreased availability of the TGFβ receptor complex results in the absence of Smad2 and Smad3 activation, thereby limiting EndMT induction.

FGF2 induces miR-20a and inhibits EndMT

Endothelial mitogens inhibit EndMT (Medici and Kalluri, 2012; Okayama et al., 2012; Ichise et al., 2014; Zhang et al., 2015). We questioned whether this inhibitory effect might be (in part) due to the induction of miR-20a. Endothelial cells treated with TGFβ1 decreased the expression of miR-20a (2.6-fold, P<0.01, Fig. 4A). All endothelial mitogens tended (P<0.1) to increase the expression of miR-20a in TGFβ1-treated endothelial cells; however, FGF2 increased miR-20a expression to a level above that of non-treated endothelial cells (2.5-fold versus control, 6.5-fold versus TGFβ1 treatment, P<0.01; Fig. 4A).

Fig. 4.

Endothelial FGF2 signaling through RAS induces microRNA-20a expression and inhibits EndMT. (A) Endothelial cells were stimulated with TGFβ1 and endothelial mitogens and assayed for miR-20a expression (n=5 per group). The addition of FGF2 increases miR-20a expression to above the level of controls. miR-20a expression is dependent on Ras and PI3K signaling (n=5 per group), as the addition of inhibitor to Ras (FTS, 5 µM) or PI3K (LY294002, 10 µM) blocks the FGF2-induced expression of miR-20a. Blockage of downstream mediators of Ras and PI3K, JNK (SP600125, 1 µM) and Erk1/2 (U0126, 5 µM) also blocks the FGF2-induced expression of miR-20a. (B–E) Light microscopy images of endothelial cells. (B) Untreated cells. Treatment of endothelial cells with TGFβ1 (C) induces hypertrophy, which is counteracted by FGF2 (E). Treatment of endothelial cells with FGF2 did not alter cellular morphology (D).Immunofluorescence analysis of VE-cadherin (n=5 per group, F–I) and SM22α (n=5 per group, K–N). (F,K) Untreated cells. TGFβ1 treatment of endothelial cells decreases VE-cadherin expression (G) and induces SM22α expression (L). Treatment of HUVECs with FGF2 maintains VE-cadherin expression (I) and inhibits SM22α expression (N) in TGFβ1-treated endothelial cells. Treatment of endothelial cells with FGF2 alone does not alter VE-cadherin (H) or SM22α expression (M). (J,O) Quantification of immunofluorescence analysis (n=5 per group). (P) Gene expression data on the mesenchymal transcription factors Snai1, Snai2 and Twist1. Expression is induced by TGFβ and inhibited by the addition of FGF2 (n=5 per group). (Q,R) Representative immunoblots of components of endothelial TGFβ signaling (Q) and their quantification (n=5 per group, R). TGFβ1 increases expression of the TGFβ receptor complex and activates Smad2 and Smad3. These expression and activity changes are inhibited by the supplementation of FGF2 to endothelial cells. Graphical results are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001 (one-way ANOVA followed by Bonferroni post-tests).

Fig. 4.

Endothelial FGF2 signaling through RAS induces microRNA-20a expression and inhibits EndMT. (A) Endothelial cells were stimulated with TGFβ1 and endothelial mitogens and assayed for miR-20a expression (n=5 per group). The addition of FGF2 increases miR-20a expression to above the level of controls. miR-20a expression is dependent on Ras and PI3K signaling (n=5 per group), as the addition of inhibitor to Ras (FTS, 5 µM) or PI3K (LY294002, 10 µM) blocks the FGF2-induced expression of miR-20a. Blockage of downstream mediators of Ras and PI3K, JNK (SP600125, 1 µM) and Erk1/2 (U0126, 5 µM) also blocks the FGF2-induced expression of miR-20a. (B–E) Light microscopy images of endothelial cells. (B) Untreated cells. Treatment of endothelial cells with TGFβ1 (C) induces hypertrophy, which is counteracted by FGF2 (E). Treatment of endothelial cells with FGF2 did not alter cellular morphology (D).Immunofluorescence analysis of VE-cadherin (n=5 per group, F–I) and SM22α (n=5 per group, K–N). (F,K) Untreated cells. TGFβ1 treatment of endothelial cells decreases VE-cadherin expression (G) and induces SM22α expression (L). Treatment of HUVECs with FGF2 maintains VE-cadherin expression (I) and inhibits SM22α expression (N) in TGFβ1-treated endothelial cells. Treatment of endothelial cells with FGF2 alone does not alter VE-cadherin (H) or SM22α expression (M). (J,O) Quantification of immunofluorescence analysis (n=5 per group). (P) Gene expression data on the mesenchymal transcription factors Snai1, Snai2 and Twist1. Expression is induced by TGFβ and inhibited by the addition of FGF2 (n=5 per group). (Q,R) Representative immunoblots of components of endothelial TGFβ signaling (Q) and their quantification (n=5 per group, R). TGFβ1 increases expression of the TGFβ receptor complex and activates Smad2 and Smad3. These expression and activity changes are inhibited by the supplementation of FGF2 to endothelial cells. Graphical results are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001 (one-way ANOVA followed by Bonferroni post-tests).

We investigated through which molecular pathway FGF2 induced the expression of miR-20a using small-molecule inhibitors to the downstream mediators of FGF2 signaling (Fig. 4A). Inhibition of Ras (4.3-fold decrease, P<0.001), phosphoinositide 3-kinase (PI3K) (3.4-fold decrease, P<0.01), Erk1 and Erk2 (Erk1/2, 7.4-fold decrease, P<0.001) and JNK1–JNK3 (7.1-fold decrease) all abrogated the FGF2-induced expression of miR-20a, whereas inhibition of the p38 MAPK family and phospholipase C (PLC) had no effect of miR-20a expression in FGF2- and TGFβ1-treated endothelial cells. As Erk1/2 and JNK are both downstream mediators of Ras signaling in endothelial cells, these data indicate that FGF2 induces the expression of miR-20a through Ras signaling.

Combined treatment of endothelial cells with TGFβ1 and FGF2 inhibited EndMT induction by TGFβ1, as indicated by the absence of cellular hypertrophy (Fig. 4B–E) and maintenance of VE-cadherin expression (Fig. 4F–I), although VE-cadherin expression was reduced compared to FGF2-treated control cells (P<0.01, Fig. 4J). Notably, FGF2 treatment abolished the ability of TGFβ1 to induce the expression of SM22α in endothelial cells (Fig. 4K–O). TGFβ-induced EndMT, which is associated with increased gene expression of the mesenchymal transcription factor factors Snai1, Snai2 and Twist1, was abrogated by the addition of FGF2 (all P<0.001; Fig. 4P).

We analyzed the expression levels of proteins involved in canonical TGFβ signaling in TGFβ1 and FGF2-treated endothelial cells. FGF2 signaling inhibited the TGFβ1-induced increase in protein expression of ALK5 (3.7-fold reduction, P<0.001), TGFBR2 (3.3-fold reduction, P<0.001) and SARA (2.3-fold reduction, P<0.05), which were expressed at levels equal to those in untreated endothelial cells (Fig. 4Q,R). Endogenous levels of ALK5 decreased (2-fold, P<0.05) in endothelial cells treated with FGF2, whereas expression levels of TGFBR2 and SARA were not affected. Decreased receptor availability in TGFβ1- and FGF2-treated endothelial cells limited the endothelial ability to activate Smad2 (2.2-fold reduction, P<0.001) and Smad3 (2.4-fold reduction, P<0.01; Fig. 4Q,R).

These data suggest that FGF2 inhibits TGFβ1-induced EndMT in part by the Ras-dependent expression of miR-20a and the subsequent reduction in TGFβ receptor complex formation.

microRNA-20a loss-of-function inhibits the FGF2-mediated protection from EndMT

To address whether the induction of miR-20a expression is required for FGF2-mediated protection from EndMT, we inhibited its expression using anti-miRs (Stenvang et al., 2012). Anti-miR-20a decreased miR-20a levels 2.5-fold (P<0.05) compared to untreated endothelial cells and 8.8-fold (P<0.001) compared to endothelial cells treated with both TGFβ and FGF2, reducing miR-20a expression to the level of TGFβ-treated cells (Fig. 5A).

Fig. 5.

MicroRNA-20a loss-of-function inhibits FGF2-mediated protection from EndMT. (A) Endothelial cells were transfected with anti-miR-20a oligonucleotides or scrambled control (indicated by – in the anti-miR-20a labels) sequences and challenged or not with TGFβ1 and FGF2 (n=5 per group). The addition of FGF2 increases miR-20a expression to above the level of controls. The addition of anti-miR-20a oligonucleotides blocks the FGF2-induced expression of miR-20a. (B–E) Light microscopy images of endothelial cells. (B) Untreated cells. Treatment of endothelial cells with TGFβ1 (C) induces hypertrophy, which is counteracted by FGF2 (D). Treatment of endothelial cells with anti-miR-20a in the presence of TGFβ and FGF2 alters cellular morphology resulting in hypertrophy (E). Immunofluorescence analysis of VE-cadherin (n=5 per group, F–I) and SM22α (n=5 per group, K–N) expression. (F,K) Untreated cells. TGFβ1 treatment of endothelial cells decreases VE-cadherin expression (G) and induces SM22α expression (L). Treatment of HUVEC with FGF2 maintains VE-cadherin expression (H) and inhibits SM22α expression (M) in TGFβ1-treated endothelial cells. Treatment of endothelial cells with anti-miR-20a in combination with TGFβ and FGF2 decreases VE-cadherin (I) and induces SM22α expression (N). (J,O) Quantification of immunofluorescence analysis (n=5 per group). (P) Gene expression data on the mesenchymal transcription factors Snai1, Snai2 and Twist1 (n=5 per group). Expression of Snai1, Snai2 and Twist1 is induced by TGFβ and inhibited by the addition of FGF2. miR-20a loss-of-function restores the expression of these proteins back to the level of that after TGFβ treatment. (Q,R) Representative immunoblots of components of endothelial TGFβ signaling (Q) and their quantification (n=5 per group, R). TGFβ1 increases expression of the TGFβ receptor complex and activates Smad2 and Smad3. These expression and activity changes are inhibited by the supplementation of FGF2 to endothelial cells. The loss-of-function of miR-20a abolishes the protective effects of FGF2. Graphical results are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001 (one-way ANOVA followed by Bonferroni post-tests).

Fig. 5.

MicroRNA-20a loss-of-function inhibits FGF2-mediated protection from EndMT. (A) Endothelial cells were transfected with anti-miR-20a oligonucleotides or scrambled control (indicated by – in the anti-miR-20a labels) sequences and challenged or not with TGFβ1 and FGF2 (n=5 per group). The addition of FGF2 increases miR-20a expression to above the level of controls. The addition of anti-miR-20a oligonucleotides blocks the FGF2-induced expression of miR-20a. (B–E) Light microscopy images of endothelial cells. (B) Untreated cells. Treatment of endothelial cells with TGFβ1 (C) induces hypertrophy, which is counteracted by FGF2 (D). Treatment of endothelial cells with anti-miR-20a in the presence of TGFβ and FGF2 alters cellular morphology resulting in hypertrophy (E). Immunofluorescence analysis of VE-cadherin (n=5 per group, F–I) and SM22α (n=5 per group, K–N) expression. (F,K) Untreated cells. TGFβ1 treatment of endothelial cells decreases VE-cadherin expression (G) and induces SM22α expression (L). Treatment of HUVEC with FGF2 maintains VE-cadherin expression (H) and inhibits SM22α expression (M) in TGFβ1-treated endothelial cells. Treatment of endothelial cells with anti-miR-20a in combination with TGFβ and FGF2 decreases VE-cadherin (I) and induces SM22α expression (N). (J,O) Quantification of immunofluorescence analysis (n=5 per group). (P) Gene expression data on the mesenchymal transcription factors Snai1, Snai2 and Twist1 (n=5 per group). Expression of Snai1, Snai2 and Twist1 is induced by TGFβ and inhibited by the addition of FGF2. miR-20a loss-of-function restores the expression of these proteins back to the level of that after TGFβ treatment. (Q,R) Representative immunoblots of components of endothelial TGFβ signaling (Q) and their quantification (n=5 per group, R). TGFβ1 increases expression of the TGFβ receptor complex and activates Smad2 and Smad3. These expression and activity changes are inhibited by the supplementation of FGF2 to endothelial cells. The loss-of-function of miR-20a abolishes the protective effects of FGF2. Graphical results are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001 (one-way ANOVA followed by Bonferroni post-tests).

The loss of miR-20a was associated with morphological changes (Fig. 5C,E) loss of VE-cadherin expression (2.8-fold decrease, P<0.001; Fig. 5G,I) and induction of SM22α expression (7.0-fold increase, P<0.001; Fig. 5L,N) compared to endothelial cells treated with both TGFβ and FGF2s (Fig. 5D,H,M). Moreover, anti-miR-20a-expressing cells had similar levels of Snai1, Snai2 and Twist1 mRNAs to TGFβ-treated cells, despite the presence of FGF2 (Fig. 5P).

We analyzed the expression levels of proteins involved in canonical TGFβ signaling in TGFβ1- and FGF2-treated endothelial cells, wherein FGF2 signaling inhibits the TGFβ1-induced increase in expression of ALK5, TGFBR2 and SARA (Fig. 5Q,R). The loss of miR-20a expression in cells treated with both TGFβ and FGF2 increased the expression of ALK5 (3.5-fold, P<0.01), TGFBR2 (1.9-fold, P<0.05) and SARA (3.0-fold, P<0.001) to the level of TGFβ-treated cells (Fig. 5Q,R).

Thus, FGF2 inhibits TGFβ-induced EndMT through the induction of miR-20a expression. miR-20a limits the expression of the TGFβ receptor complex (i.e. ALK5, TGFBR2 and SARA). The loss of miR-20a expression is associated with increased Smad2 and Smad3 activity, and thus the loss of the FGF2-mediated protection from EndMT.

MicroRNA-20a inhibits the induction, but not the progression of TGFβ-mediated EndMT

We questioned whether miR-20a was able to reverse TGFβ-induced EndMT by transfecting endothelial cells with miR-20a mimics 72 h after TGFβ treatment (Fig. 6A). Interestingly, miR-20a gain-of-function limited the number of endothelial cells undergoing EndMT, as indicated by the increase in cells that expressed VE-cadherin (Fig. 6E–H) and decrease in the number of cells that express SM22α (Fig. 6J–M) compared to TGFβ-treated cells (all P<0.001). However, in cells that had dim expression of VE-cadherin, miR-20a gain-of-function did not increase VE-cadherin expression to above that of TGFβ-stimulated cells (EndMT cells) and the expression of VE-cadherin in the highly expressing population was not different from untreated endothelial cells (Fig. 6I). Similarly, SM22α expression in the dim population was similar to untreated, whereas SM22α expression in the highly expressing population was indistinguishable from TGFβ-treated cells (Fig. 6N). These data imply that miR-20a reduces the number of cells entering EndMT, but does not affect TGFβ signaling in cells that have already entered the EndMT program.

Fig. 6.

MicroRNA-20a inhibits the induction but not the progression of TGFβ-mediated EndMT. (A) Schematic of the experimental procedure. Endothelial cells were treated with TGFβ for 3 days and subsequently transfected with miR-20a mimics or scrambled control sequences (n=4 per group). (B–D) Cells uninterruptedly treated with TGFβ showed signs of hypertrophy (C), whereas cells with gain-of-function for miR-20a had a diverse phenotype, i.e. some cells were hypertrophic and others were not (D). Untreated cells (B). (E–H) TGFβ treatment decreased VE-cadherin expression (F), whereas miR-20a inhibited this decrease in ∼50% of the cells (G). A quantification is shown in H (n=4 per group). (E) Untreated cells. (J–M) SM22α expression was induced by TGFβ treatment (K) and reduced by the addition of miR-20a (L). Untreated cells (J). A quantification is shown in M (n=4 per group). (I,N) Cells that expressed VE-cadherin after treatment with both TGFβ and miR-20a mimics (high), expressed VE-cadherin at levels similar to native endothelial cells, whereas dim cells expressed VE-cadherin at similar levels to TGFβ-treated cells (n=4 per group, I). Cells that expressed SM22α after treatment with TGFβ and miR-20a mimics (high), expressed SM22α at levels similar to cells that had undergone EndMT, whereas dim cells expressed SM22α at similar levels to native endothelial cells (n=4 per group, N). Graphical results are mean±s.e.m. ***P<0.001; ns, not significant (one-way ANOVA followed by Bonferroni post-tests).

Fig. 6.

MicroRNA-20a inhibits the induction but not the progression of TGFβ-mediated EndMT. (A) Schematic of the experimental procedure. Endothelial cells were treated with TGFβ for 3 days and subsequently transfected with miR-20a mimics or scrambled control sequences (n=4 per group). (B–D) Cells uninterruptedly treated with TGFβ showed signs of hypertrophy (C), whereas cells with gain-of-function for miR-20a had a diverse phenotype, i.e. some cells were hypertrophic and others were not (D). Untreated cells (B). (E–H) TGFβ treatment decreased VE-cadherin expression (F), whereas miR-20a inhibited this decrease in ∼50% of the cells (G). A quantification is shown in H (n=4 per group). (E) Untreated cells. (J–M) SM22α expression was induced by TGFβ treatment (K) and reduced by the addition of miR-20a (L). Untreated cells (J). A quantification is shown in M (n=4 per group). (I,N) Cells that expressed VE-cadherin after treatment with both TGFβ and miR-20a mimics (high), expressed VE-cadherin at levels similar to native endothelial cells, whereas dim cells expressed VE-cadherin at similar levels to TGFβ-treated cells (n=4 per group, I). Cells that expressed SM22α after treatment with TGFβ and miR-20a mimics (high), expressed SM22α at levels similar to cells that had undergone EndMT, whereas dim cells expressed SM22α at similar levels to native endothelial cells (n=4 per group, N). Graphical results are mean±s.e.m. ***P<0.001; ns, not significant (one-way ANOVA followed by Bonferroni post-tests).

DISCUSSION

Here, we show that FGF2-induced activation of Ras signaling induces the expression of miR-20a. MiR-20a targets multiple proteins in the TGFβ receptor complex, namely ALK5 (TGFBR1), TGFBR2 and SARA, thereby reducing their expression levels and inhibiting TGFβ1-induced activation of Smad2 and Smad3 and the resulting expression of mesenchymal genes. Corroboratively, endothelial cells show reduced susceptibility to TGFβ1 and maintain expression of endothelial cell markers and functions. Remarkably, when administered 3 days post TGFβ treatment, miR-20a limited the number of cells that entered EndMT, but did not affect cells that had already entered the EndMT program. Importantly, our data implies that miR-20a is a new mediator of endothelial TGFβ1 responsiveness and EndMT.

Postnatal EndMT is associated with fibroproliferative diseases, such as cancer progression and metastasis (Potenta et al., 2008), cardiac (Zeisberg et al., 2007) and kidney fibrosis (Zeisberg et al., 2008), and has received a large research interest over the past decade. TGFβ signaling is recognized as the key driving force of EndMT (Derynck and Akhurst, 2007; Pannu et al., 2007; van Meeteren and ten Dijke, 2012) and inhibitors of TGFβ signaling, such as BMP7 (Zeisberg et al., 2007), receive increasing therapeutic interest as anti-fibrotic compounds. However, little is known about endogenous inhibitors of TGFβ signaling in endothelial cells during EndMT.

MicroRNAs are involved in EndMT (Bijkerk et al., 2012; Fang and Davies, 2012; Ghosh et al., 2012; Kumarswamy et al., 2012; Zhang et al., 2013) and fibroproliferative diseases (Thum et al., 2008; van Rooij et al., 2008; Kato et al., 2009; Chung et al., 2010) by targeting genes that are crucial for endothelial homeostasis or activating fibroblasts. However, the role of microRNAs that are lost during EndMT and fibroproliferative diseases remains elusive. Results from the present study indicate that miR-20a plays an important role in the endothelial susceptibility to TGFβ by actively suppressing expression of components of the TGFβ receptor complex in TGFβ-stimulated cells. Hence, in order to react to and propagate TGFβ signals in endothelial cell, miR-20a needs to decrease. Indeed, TGFβ1-stimulated endothelial cells had a reduced expression of miR-20a. Notably, endothelial cells transfected with miR-20a mimics had a reduced sensitivity to TGFβ1 and were resistant to EndMT. Interestingly, miR-20a mimics efficiently inhibited ALK5, TGFBR2 and SARA expression in TGFβ-treated cells, but did not affect the expression of these proteins in native endothelial cells. As the efficacy of a microRNA in decreasing the expression its target genes relies on the expression level not only of the target gene under investigation, but also on the presence of other mRNAs that are targeted by the same microRNA [known as competing endogenous RNAs; ceRNAs (Salmena et al., 2011)], our data might indicate the presence of ceRNAs in unstimulated endothelial cells, which inhibits the decrease in ALK5, TGFBR2 and SARA.

In fibroproliferative diseases, the endothelium shows resistance to a number of mitogens, such as FGF2, vascular endothelial growth factor A (VEGFa) and insulin-like growth factor-1 (IGF1) (Baelde et al., 2007; Cheng et al., 2014; Stiedl et al., 2015). We therefore wondered whether these mitogens could relay some protective effects at non-affected sited versus affected sites where the response to these mitogens is lost. Hence, we assessed whether these mitogens might induce the expression of miR-20a and counter TGFβ-induced EndMT. Indeed, in healthy endothelial cells, FGF2 strongly induces the expression of miR-20a in a Ras-dependent manner, reducing the endothelial expression of the TGFβ receptor complex, and prevents EndMT induction by TGFβ1. This implies that therapeutic induction of miR-20a expression or the targeted delivery of miR-20a mimics might pose a novel therapy to treat EndMT in fibroproliferative diseases. Thus, miR-20a links two crucial growth factor signaling pathways in endothelial homeostasis and dysfunction (Fig. 7). Recent investigations using endothelial-specific deletion of FGF receptor 1 (FGFR1) (Cheng et al., 2014) or FGF receptor substrate 2α (Frs2α) (Chen et al., 2015), corroborate the importance of FGF2 signaling in the inhibition of EndMT. The loss of FGFR1 activity aggravates atherosclerosis development and EndMT progression, in part by increased Smad2 activity (Chen et al., 2014; Chen et al., 2015). Moreover, FGFR1 expression is highly reduced during human atherosclerosis development, a process characterized by EndMT (Chen et al., 2015). Interestingly, the expression of another microRNA that inhibits ALK5 expression, let-7b, is also highly dependent on FGF2 signaling, as the loss of FGFR1 results in the ablation of let-7b levels and an increase TGFβ activity and EndMT (Chen et al., 2012). These data corroborate and extend earlier reports where FGF2 was reported to reduce endothelial sensitivity to TGFβ (Kawai-Kowase et al., 2004; Ichise et al., 2014).

Fig. 7.

Schematic representation of the FGF2-mediated regulation of the endothelial TGFβ1 susceptibility by controlling miR-20a levels. (A) In the absence of FGF2 signaling, TGFβ1 binds to its heteromeric receptor formed by ALK5 and TGFBR2. SARA recruits the R-Smads to the ALK5 kinase. Activation of Smad2 and Smad3, and concomitant complexation with Smad4, results in nuclear translocation and transcription of mesenchymal genes and repression of endothelial genes. (B) FGF2 signaling activates its downstream mediators Ras and PI3K, resulting in the expression of miR-20a. miR-20a represses the expression of the TGFβ receptor complex (ALK5 and TGFBR2) and SARA, thus inhibiting Smad2 and Smad3 activation and the culminating induction of EndMT.

Fig. 7.

Schematic representation of the FGF2-mediated regulation of the endothelial TGFβ1 susceptibility by controlling miR-20a levels. (A) In the absence of FGF2 signaling, TGFβ1 binds to its heteromeric receptor formed by ALK5 and TGFBR2. SARA recruits the R-Smads to the ALK5 kinase. Activation of Smad2 and Smad3, and concomitant complexation with Smad4, results in nuclear translocation and transcription of mesenchymal genes and repression of endothelial genes. (B) FGF2 signaling activates its downstream mediators Ras and PI3K, resulting in the expression of miR-20a. miR-20a represses the expression of the TGFβ receptor complex (ALK5 and TGFBR2) and SARA, thus inhibiting Smad2 and Smad3 activation and the culminating induction of EndMT.

Interestingly, although we and others describe a protective effect of FGF2 against EndMT induction in mature and progenitor endothelial cells (Moonen et al., 2010; Ichise et al., 2014), others have linked FGF2 signaling to EndMT induction (Ghosh et al., 2010; Lee et al., 2012). Indeed, FGF2 is known to inhibit EndMT in macrovascular endothelial cells, whereas it induces EndMT is some microvascular endothelial cell types (Lee et al., 2004; Lee and Kay, 2006). Whether this difference in EndMT induction is dependent on the organ-of-origin, as not all microvascular endothelial cells undergo EndMT following FGF2 stimulation (Ichise et al., 2014), or is derived from some microenvironmental cue is currently unknown and warrants more investigation.

Moreover, although multiple receptor tyrosine kinases (RTKs) are implemented in the protection against EndMT (Medici et al., 2010; Okayama et al., 2012), only FGF2 signaling induced miR-20a. RTKs depend on scaffold molecules that organize their downstream effects. Therefore, it is appealing to hypothesize that FGF2 utilizes a unique scaffold protein for its signaling, distinct from other RTKs. Whether FGF2 functions through a unique scaffold protein is currently unknown and warrants further investigation.

In summary, FGF2 regulates endothelial TGFβ1 signaling by controlling ALK5, TGFBR2 and SARA expression, through control of miR-20a levels. Loss of FGF2 signaling combined with a TGFβ1 challenge decreases miR-20a levels and increases endothelial responsiveness to TGFβ1 through elevated receptor complex levels and activation of Smad2 and Smad3. TGFβ1 treatment culminates in EndMT, which is abrogated by either FGF2 or exogenous miR-20a. These data suggest that miR-20a is a new regulator of endothelial TGFβ signaling and might be a new target for therapeutic silencing of TGFβ activity in fibroproliferative diseases.

MATERIALS AND METHODS

Endothelial cell culture, transfection and stimulations

Human umbilical vein endothelial cells (HUVECs; Lonza, Verviers, Belgium) were cultured in gelatin-coated culture flasks in endothelial cell medium containing RPMI 1640 (Lonza, Verviers, Belgium) supplemented with 20% fetal bovine serum (FBS; Life Technologies, Carlsbad, CA), 50 µg/ml bovine pituitary extract (Lonza, Verviers, Belgium), 2 mM L-glutamine, 1% penicillin-streptomycin (both Sigma-Aldrich, St Louis, MO) and 5 U/ml heparin (Leo Pharma, Amsterdam, The Netherlands). HUVECs were used between passages 3 and 6.

HUVECs were stimulated with TGFβ1, FGF2, hepatocyte growth factor (HGF), IGF1 or VEGFa (all 10 ng/ml, all Peprotech, Rocky Hill, NJ) for 72 h. Small-molecule inhibitors to ALK5 (SB431542, 5 µM), p38 MAPK (SB203580, 1 µM, both Sigma-Aldrich), Ras [farnesyl thiosalicylic acid (FTS); 5 µM, Cayman Chemical, Ann Arbor, MI], PI3K (LY294002, 10 µM, SelleckChem, Munich, Germany), Erk1/2 (U0126, 5 µM, Promega, Madison, WI) or JNK (SP600125, 1 µM, EMD Millipore, Darmstadt, Germany) were used where indicated for 72 h.

HUVECs were transfected with microRNA-20a (PM10057), anti-miR-20a (AM10057) or a scrambled control sequence (AM17110, both 100 nM, all AMBION/Life Technologies, Carlsbad, CA), using the siRNA Reagent System (Santa Cruz Biotechnology, Heidelberg, Germany) according to manufacturer's instructions. After overnight incubation, HUVECs were challenged with the appropriate stimuli.

3′UTR reporter assays

Isolation of 3′UTR fragments was performed from a cDNA pool derived from various human tissues using specific primers for the ALK5 3′UTR (sense 5′-TTCTACAGCTTTGCCTGAAC-3′, antisense 5′-GTCTGGGAATGTCTTTAATT-3′), the TGFBR2 3′UTR (sense 5′-CTCTTCTGGGGCAGGCTGGG-3′, antisense 5′-AGCTACTAGGAATGGGAACAG-3′), the SARA 3′UTR (sense 5′-ACAGAGAAGACTTCATTTTT-3′, antisense 5′-CAGTGTGGAATTATCCTTTT-3′), the SMAD2 3′UTR (sense 5′-AGCTTCACCAATCAAGTCCC-3′, antisense 5′-AACATGGTAAACAACTCAAA-3′), the SMAD3 3′UTR (sense 5′-AGACATCAAGTATGGTAGGG-3′, antisense 5′-CAGACTGAGCTCCTGGCACA-3′) or the SMAD4 3′UTR (sense 5′-GGTCTTTTACCGTTGGGGCC-3′, antisense 5′-AGTTGGCTTTCTCTTTTAAT-3′) (all 0.6 µM, Biologio, Leiden, The Netherlands). Sense and antisense primers were extended with SgfI (GCGATCGC) and NotI (GCGGCCGC) restriction sequences, respectively. Amplification was performed using the DyNAzyme EXT PCR kit (Finnzymes, Vantaa, Finland) according to the manufacturer's instructions. Amplicon size was validated by gel electrophoresis on 1% agarose gels.

Amplicons were purified using the QIAquick PCR Purification kit (Qiagen, Venlo, The Netherlands) and cloned into the dual luciferase reporter vector psiCHECK-2 (Promega, Madison, WI) using T4 DNA Ligase (Fermentas/Thermo Fisher Sci., Waltham, MA) according to standard protocols.

HEK293 cells (Sigma-Aldrich, St. Louis, MO) were maintained in DMEM (Lonza, Verviers, Belgium) containing 10% FBS, 2 mM L-glutamine and 1% Penicillin/Streptomycin. HEK293 cells were transfected with 100 ng/ml UTR reporter plasmid and 50 nM miR-20a mimic or scrambled control (Ambion/Life Technologies, Carlsbad, CA) using Endofectin (GeneCopoeia, Rockville, MD). 48 h post-transfection, luciferase activity was assayed using the DualGlo Luciferase assay system (Promega, Madison, WI) and recorded for 1 s on a Luminoskan ASCENT (Thermo Scientific, Waltham, MA) according to manufacturer's instructions.

MicroRNA and mRNA expression analysis

Total RNA was isolated using TRIzol Reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. RNA quantity and purity were assessed by spectrophotometric analysis (Nanodrop, Wilmington, DE) wherein both the ratios of absorbance at 260 nm to that at 230 nm (A260/230) and absorbance at 260 nm to that at 280 nm (A260/280) were >1.8. RNA integrity was assessed by gel electrophoresis on a 2% agarose gel. For microRNA expression analysis, 20 ng total RNA was reversely transcribed using the Taqman MicroRNA Reverse Transcription kit (Applied Biosystems, Carlsbad, CA) using specific miR-20a (5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACCTACCTGC-3′) or RNU6B (U6; 5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACAA AAATATGG-3′) stem loop primers. Quantitative PCR expression analysis was performed on a reaction mixture containing 10 ng cDNA equivalent, 0.5 µM miR primers (miR-20a; 5′-TGCGGTTAAAGTGCTTATAGT-3′, RNU6B; 5′-TGCGGCTGCGCAAGGATGA-3′, antisense 5′-GTGCAGGGTCCGAGGT-3′, all Biolegio, Leiden, The Netherlands) and FastStart SYBR Green (Roche, Almere, The Netherlands). For mRNA expression analysis, 500 ng total RNA was reversely transcribed using the RevertAid First Strand cDNA Synthesis Kit (Applied Biosystems, Carlsbad, CA) using specific Snai1 (sense 5′-GCTGCAGGACTCTAATCCAGA-3′; antisense 5′-ATCTCCGGAGGTGGGATG-3′), Snai2 (sense 5′-TGGTTGCTTCAAGGACACAT-3′; antisense 5′-GTTGCAGT-GAGGGCAAGAA-3′), Twist1 (sense 5′-AAGGCATCACT-ATGGACTTTCTCT-3′; antisense 5′-GCCAGTTTGATCCCAGTATTTT-3′) and GAPDH (sense 5′-AGCCACATCGCTCAGACAC-3′; antisense 5′-GCCCAATACGACCAAATCC-3′) primers. Quantitative PCR expression analysis was performed on a reaction mixture containing 10 ng cDNA equivalent, 0.5 µM sense primers and 0.5 µM antisense primers (all Biolegio, Leiden, The Netherlands) and FastStart SYBR Green (Roche, Almere, The Netherlands). Analyses were run on a Viia7 real-time PCR system (Applied Biosystems, Carlsbad, CA).

Immunofluorescence

Cells were fixed using 4% paraformaldehyde in PBS at room temperature for 15 min. For intracellular staining, fixed cells were permeabilized using 0.5% Triton X-100 in PBS (Sigma-Aldrich) at room temperature for 10 min. Blocking of specific antibody activity was performed using 2% bovine serum albumin (BSA) in PBS for 10 min. Samples were incubated with antibodies to VE-cadherin (1:200, Cell Signaling #2500, Danvers, MA) or SM22α (1:250, Abcam #14106, Cambridge, UK) in PBS containing 2% BSA at 4°C overnight. Samples were washed extensively with 0.05% Tween-20 in PBS and incubated with Alexa-Fluor-594-conjugated antibodies to rabbit IgG (Life Technologies, Carlsbad, CA, #A21207) in DAPI with PBS with 2% BSA at room temperature for 1 h. Image analysis was performed on TissueFAXS (TissueGnostics, Vienna, Austria) in fluorescence mode, in combination Zeiss AxioObserver Z1 microscope. Data analysis was performed using TissueQuest fluorescence (TissueGnostics, Vienna, Austria) software. For all immunofluorescence analyses, 1000–2000 individual cells were analyzed.

Immunoblotting

Cells were lysed in RIPA buffer (Thermo Scientific, Waltham, MA) supplemented with 0.1% proteinase inhibitor cocktail (Sigma-Aldrich). Samples (30 µg/lane) were loaded on 10% SDS-PAGE gels and blotted onto nitrocellulose membranes. Membranes were incubated with antibodies to ALK5 (1:500, #31013), TGFBR2 (1:500, #61213), SARA (1:1000, #124875, all Abcam, Cambridge, UK), pSmad2 (Ser465 and Ser467, 1:500, #3108), pSmad3 (Ser423 and Ser 425, 1:500, #9520), Smad2 and Smad3 (1:500, #3102, all Cell Signaling, Danvers, MA) and GAPDH (1:2000, #9485, Abcam, Cambridge, UK) in Odyssey Blocking Buffer overnight at 4°C. Next, membranes were incubated with IRDye680-conjugated antibodies to rabbit IgG (LI-COR Biosciences, Bad Homburg, Germany) diluted 1:10,000 in Odyssey Blocking Buffer at room temperature for 1 h. Proteins were visualized using the Odyssey® Infrared Imaging System (LI-COR Bioscience). Densitometric analysis was performed using TotalLab TL120 1D (Nonlinear Dynamics, Durham, NC).

Matrigel sprouting

10 µl of MatriGel (BD Biosciences, San Jose, CA) was solidified in µ-Slide Angiogenesis (Ibidi, Martinsried, Germany). 15,000 cells per well were cultured on the solidified MatriGel in endothelial growth medium, overnight. Formation of sprouts was analyzed by conventional light microscopic analysis.

Statistical analysis

All experimental data were obtained from at least five independent experiments. Data is expressed as mean±s.e.m. Data were analyzed using one-way ANOVA followed by Bonferroni post-tests. P<0.05 was considered to be statistically significant.

Acknowledgements

We acknowledge Mr. K. Sjollema (UMCG, UMIC) for expert technical assistance with fluorescence imaging. Imaging was performed at the UMCG Imaging Center (UMIC), supported by the Netherlands Organization for Health Research and Development (ZonMW grant 40-00506-98-9021).

Footnotes

Author contributions

G.K. and J.-R.A.J.M. conceived and coordinated the study. G.K. and A.C.P.C. wrote the paper. G.K., A.C.P.C. and M.G.L.B. designed, performed and analyzed the experiments shown in Figs 16. G.K. and A.C.P.C. prepared the figures. All authors reviewed the results and approved the final version of the manuscript.

Funding

This work was supported by the Groningen University Institute for Drug Exploration (GUIDE) (to G.K.); and a ZonMW and NWO Innovational Research Incentive grant [grant number 916.11.022 to G.K].

References

Baelde
,
H. J.
,
Eikmans
,
M.
,
Lappin
,
D. W. P.
,
Doran
,
P. P.
,
Hohenadel
,
D.
,
Brinkkoetter
,
P.-T.
,
van der Woude
,
F. J.
,
Waldherr
,
R.
,
Rabelink
,
T. J.
,
de Heer
,
E.
, et al. 
(
2007
).
Reduction of VEGF-A and CTGF expression in diabetic nephropathy is associated with podocyte loss
.
Kidney Int.
71
,
637
-
645
.
Betel
,
D.
,
Wilson
,
M.
,
Gabow
,
A.
,
Marks
,
D. S.
and
Sander
,
C.
(
2008
).
The microRNA.org resource: targets and expression
.
Nucl. Acids Res.
36
,
D149
-
D153
.
Betel
,
D.
,
Koppal
,
A.
,
Agius
,
P.
,
Sander
,
C.
and
Leslie
,
C.
(
2010
).
Comprehensive modeling of microRNA targets predicts functional non-conserved and non-canonical sites
.
Genome Biol.
11
,
R90
.
Bijkerk
,
R.
,
de Bruin
,
R. G.
,
van Solingen
,
C.
,
Duijs
,
J. M. G. J.
,
Kobayashi
,
K.
,
van der Veer
,
E. P.
,
ten Dijke
,
P.
,
Rabelink
,
T. J.
,
Goumans
,
M. J.
and
van Zonneveld
,
A. J.
(
2012
).
MicroRNA-155 functions as a negative regulator of RhoA signaling in TGF-β-induced endothelial to mesenchymal transition
.
MicroRNA
1
,
2
-
10
.
Chen
,
P. -Y.
,
Qin
,
L.
,
Barnes
,
C.
,
Charisse
,
K.
,
Yi
,
T.
,
Zhang
,
X.
,
Ali
,
R.
,
Medina
,
Pedro
,
P.
,
Yu
,
J.
,
Slack
, and
Frank
,
J
, et al. 
(
2012
).
FGF Regulates TGF-β Signaling and Endothelial-to-Mesenchymal Transition via Control of let-7 miRNA Expression
.
Cell Rep.
2
,
1684
-
1696
.
Chen
,
P. Y.
,
Qin
,
L.
,
Tellides
,
G.
and
Simons
,
M.
(
2014
).
Fibroblast growth factor receptor 1 is a key inhibitor of TGFbeta signaling in the endothelium
.
Sci Signal
7
,
ra90
.
Chen
,
P. Y.
,
Qin
,
L.
,
Baeyens
,
N.
,
Li
,
G.
,
Afolabi
,
T.
,
Budatha
,
M.
,
Tellides
,
G.
,
Schwartz
,
M. A.
, and
Simons
,
M.
(
2015
).
Endothelial-to-mesenchymal transition drives atherosclerosis progression
.
J. Clin. Invest.
125
,
4514
-
4528
.
Cheng
,
M. F.
,
Chen
,
L. J.
,
Wang
,
M. C.
,
Hsu
,
C. T.
and
Cheng
,
J. T.
(
2014
).
Decrease of FGF receptor (FGFR) and interstitial fibrosis in the kidney of streptozotocin-induced diabetic rats
.
Horm. Metab. Res.
46
,
1
-
7
.
Chung
,
A. C. K.
,
Huang
,
X. R.
,
Meng
,
X.
and
Lan
,
H. Y.
(
2010
).
miR-192 mediates TGF-beta/Smad3-driven renal fibrosis
.
J. Am. Soc. Nephrol.
21
,
1317
-
1325
.
Cooley
,
B. C.
,
Nevado
,
J.
,
Mellad
,
J.
,
Yang
,
D.
,
St Hilaire
,
C.
,
Negro
,
A.
,
Fang
,
F.
,
Chen
,
G.
,
San
,
H.
,
Walts
,
A. D.
, et al. 
(
2014
).
TGF-beta signaling mediates endothelial-to-mesenchymal transition (EndMT) during vein graft remodeling
.
Sci. Transl. Med.
6
,
227ra34
.
Derynck
,
R.
and
Akhurst
,
R.
(
2007
).
Differentiation plasticity regulated by TGF-β family proteins in development and disease
.
Nat. Cell Biol.
9
,
1000
-
1004
.
Fang
,
Y.
and
Davies
,
P. F.
(
2012
).
Site-specific microRNA-92a regulation of Krüppel-like factors 4 and 2 in atherosusceptible endothelium
.
Arterioscler. Thromb. Vasc. Biol.
32
,
979
-
987
.
Ghosh
,
A. K.
,
Bradham
,
W. S.
,
Gleaves
,
L. A.
,
De Taeye
,
B.
,
Murphy
,
S. B.
,
Covington
,
J. W.
and
Vaughan
,
D. E.
(
2010
).
Genetic deficiency of plasminogen activator inhibitor-1 promotes cardiac fibrosis in aged mice: involvement of constitutive transforming growth factor-β signaling and endothelial-to-mesenchymal transition
.
Circulation
122
,
1200
-
1209
.
Ghosh
,
A. K.
,
Nagpal
,
V.
,
Covington
,
J. W.
,
Michaels
,
M. A.
and
Vaughan
,
D. E.
(
2012
).
Molecular basis of cardiac endothelial-to-mesenchymal transition (EndMT): differential expression of microRNAs during EndMT
.
Cell Signal.
24
,
1031
-
1036
.
Goumans
,
M.-J.
,
van Zonneveld
,
A. J.
and
ten Dijke
,
P.
(
2008
).
Transforming growth factor beta–induced endothelial-to-mesenchymal transition: a switch to cardiac fibrosis?
Trends Cardiovasc. Med.
18
,
293
-
298
.
Grimson
,
A.
,
Farh
,
K. K.-H.
,
Johnston
,
W. K.
,
Garrett-Engele
,
P.
,
Lim
,
L. P.
and
Bartel
,
D. P.
(
2007
).
MicroRNA targeting specificity in mammals: determinants beyond seed pairing
.
Mol. Cell
27
,
91
-
105
.
Hashimoto
,
N.
,
Phan
,
S. H.
,
Imaizumi
,
K.
,
Matsuo
,
M.
,
Nakashima
,
H.
,
Kawabe
,
T.
,
Shimokata
,
K.
and
Hasegawa
,
Y.
(
2010
).
Endothelial–mesenchymal transition in bleomycin-induced pulmonary fibrosis
.
Am. J. Respir. Cell Mol. Biol.
43
,
161
-
172
.
Ichise
,
T.
,
Yoshida
,
N.
and
Ichise
,
H.
(
2014
).
FGF2-induced Ras-MAPK signalling maintains lymphatic endothelial cell identity by upregulating endothelial-cell-specific gene expression and suppressing TGFβ signalling through Smad2
.
J. Cell Sci.
127
,
845
-
857
.
Kato
,
M.
,
Putta
,
S.
,
Wang
,
M.
,
Yuan
,
H.
,
Lanting
,
L.
,
Nair
,
I.
,
Gunn
,
A.
,
Nakagawa
,
Y.
,
Shimano
,
H.
,
Todorov
,
I.
, et al. 
(
2009
).
TGF-beta activates Akt kinase through a microRNA-dependent amplifying circuit targeting PTEN
.
Nat. Cell Biol.
11
,
881
-
889
.
Kawai-Kowase
,
K.
,
Sato
,
H.
,
Oyama
,
Y.
,
Kanai
,
H.
,
Sato
,
M.
,
Doi
,
H.
and
Kurabayashi
,
M.
(
2004
).
Basic fibroblast growth factor antagonizes transforming growth factor-beta1-induced smooth muscle gene expression through extracellular signal-regulated kinase 1/2 signaling pathway activation
.
Arterioscler. Thromb. Vasc. Biol.
24
,
1384
-
1390
.
Kokudo
,
T.
,
Suzuki
,
Y.
,
Yoshimatsu
,
Y.
,
Yamazaki
,
T.
,
Watabe
,
T.
and
Miyazono
,
K.
(
2008
).
Snail is required for TGFbeta-induced endothelial-mesenchymal transition of embryonic stem cell-derived endothelial cells
.
J. Cell Sci.
121
,
3317
-
3324
.
Krenning
,
G.
,
Moonen
,
J. R.
,
van Luyn
,
M. J.
and
Harmsen
,
M. C.
(
2008
).
Vascular smooth muscle cells for use in vascular tissue engineering obtained by endothelial-to-mesenchymal transdifferentiation (EnMT) on collagen matrices
.
Biomaterials
29
,
3703
-
3711
.
Krenning
,
G.
,
Zeisberg
,
E. M.
and
Kalluri
,
R.
(
2010
).
The origin of fibroblasts and mechanism of cardiac fibrosis
.
J. Cell Physiol.
225
,
631
-
637
.
Kumarswamy
,
R.
,
Volkmann
,
I.
,
Jazbutyte
,
V.
,
Dangwal
,
S.
,
Park
,
D.-H.
and
Thum
,
T.
(
2012
).
Transforming growth factor-β-induced endothelial-to-mesenchymal transition is partly mediated by microRNA-21
.
Arterioscler. Thromb. Vasc. Biol.
32
,
361
-
369
.
Lee
,
H. T.
,
Lee
,
H. T.
,
Lee
,
J. G.
,
Na
,
M.
and
Kay
,
E. P.
(
2004
).
FGF-2 induced by interleukin-1 beta through the action of phosphatidylinositol 3-kinase mediates endothelial mesenchymal transformation in corneal endothelial cells
.
J. Biol. Chem.
279
,
32325
-
32332
.
Lee
,
J. G.
and
Kay
,
E. P.
(
2006
).
FGF-2-mediated signal transduction during endothelial mesenchymal transformation in corneal endothelial cells
.
Exp. Eye Res.
83
,
1309
-
1316
.
Lee
,
J. G.
,
Ko
,
M. K.
and
Kay
,
E. P.
(
2012
).
Endothelial mesenchymal transformation mediated by IL-1beta-induced FGF-2 in corneal endothelial cells
.
Exp. Eye Res.
95
,
35
-
39
.
Maleszewska
,
M.
,
Moonen
,
J.-R. A. J.
,
Huijkman
,
N.
,
van de Sluis
,
B.
,
Krenning
,
G.
and
Harmsen
,
M. C.
(
2013
).
IL-1beta and TGFbeta2 synergistically induce endothelial to mesenchymal transition in an NFkappaB-dependent manner
.
Immunobiology
218
,
443
-
454
.
Medici
,
D.
and
Kalluri
,
R.
(
2012
).
Endothelial–mesenchymal transition and its contribution to the emergence of stem cell phenotype
.
Semin. Cancer Biol.
22
,
379
-
384
.
Medici
,
D.
and
Olsen
,
B. R.
(
2012
).
The role of endothelial-mesenchymal transition in heterotopic ossification
.
J. Bone Miner. Res.
27
,
1619
-
1622
.
Medici
,
D.
,
Shore
,
E. M.
,
Lounev
,
V. Y.
,
Kaplan
,
F. S.
,
Kalluri
,
R.
and
Olsen
,
B. R.
(
2010
).
Conversion of vascular endothelial cells into multipotent stem-like cells
.
Nat. Med.
16
,
1400
-
1406
.
Medici
,
D.
,
Potenta
,
S.
and
Kalluri
,
R.
(
2011
).
Transforming growth factor-beta2 promotes Snail-mediated endothelial–mesenchymal transition through convergence of Smad-dependent and Smad-independent signalling
.
Biochem. J.
437
,
515
-
520
.
Moonen
,
J. R. A. J.
,
Krenning
,
G.
,
Brinker
,
M. G. L.
,
Koerts
,
J. A.
,
van Luyn
,
M. J. A.
and
Harmsen
,
M. C.
(
2010
).
Endothelial progenitor cells give rise to pro-angiogenic smooth muscle-like progeny
.
Cardiovasc. Res.
86
,
506
-
515
.
Moonen
,
J.-R. A. J.
,
Lee
,
E. S.
,
Schmidt
,
M.
,
Maleszewska
,
M.
,
Koerts
,
J. A.
,
Brouwer
,
L. A.
,
van Kooten
,
T. G.
,
van Luyn
,
M. J. A.
,
Zeebregts
,
C. J.
,
Krenning
,
G.
, et al. 
(
2015
).
Endothelial-to-mesenchymal transition contributes to fibro-proliferative vascular disease and is modulated by fluid shear stress
.
Cardiovasc. Res.
108
,
377
-
386
.
Okayama
,
K.
,
Azuma
,
J.
,
Dosaka
,
N.
,
Iekushi
,
K.
,
Sanada
,
F.
,
Kusunoki
,
H.
,
Iwabayashi
,
M.
,
Rakugi
,
H.
,
Taniyama
,
Y.
and
Morishita
,
R.
(
2012
).
Hepatocyte growth factor reduces cardiac fibrosis by inhibiting endothelial-mesenchymal transition
.
Hypertension
59
,
958
-
965
.
Pannu
,
J.
,
Nakerakanti
,
S.
,
Smith
,
E.
,
Dijke
,
P. t.
and
Trojanowska
,
M.
(
2007
).
Transforming growth Factor-beta receptor type I-dependent fibrogenic gene program is mediated via activation of Smad1 and ERK1/2 pathways
.
J. Biol. Chem.
282
,
10405
-
10413
.
Potenta
,
S.
,
Zeisberg
,
E.
and
Kalluri
,
R.
(
2008
).
The role of endothelial-to-mesenchymal transition in cancer progression
.
Br. J. Cancer
99
,
1375
-
1379
.
Qiu
,
P.
,
Feng
,
X.-H.
and
Li
,
L.
(
2003
).
Interaction of Smad3 and SRF-associated complex mediates TGF-β1 signals to regulate SM22 transcription during myofibroblast differentiation
.
J. Mol. Cell. Cardiol.
35
,
1407
-
1420
.
Quiat
,
D.
and
Olson
,
E. N.
(
2013
).
MicroRNAs in cardiovascular disease: from pathogenesis to prevention and treatment
.
J. Clin. Invest.
123
,
11
-
18
.
Salmena
,
L.
,
Poliseno
,
L.
,
Tay
,
Y.
,
Kats
,
L.
and
Pandolfi
,
P. P.
(
2011
).
A ceRNA hypothesis: the Rosetta Stone of a hidden RNA language?
Cell
146
,
353
-
358
.
Stenvang
,
J.
,
Petri
,
A.
,
Lindow
,
M.
,
Obad
,
S.
and
Kauppinen
,
S.
(
2012
).
Inhibition of microRNA function by antimiR oligonucleotides
.
Silence
3
,
1
.
Stiedl
,
P.
,
McMahon
,
R.
,
Blaas
,
L.
,
Stanek
,
V.
,
Svinka
,
J.
,
Grabner
,
B.
,
Zollner
,
G.
,
Kessler
,
S. M.
,
Claudel
,
T.
,
Müller
,
M.
, et al. 
(
2015
).
Growth hormone resistance exacerbates cholestasis-induced murine liver fibrosis
.
Hepatology
61
,
613
-
626
.
Thum
,
T.
,
Gross
,
C.
,
Fiedler
,
J.
,
Fischer
,
T.
,
Kissler
,
S.
,
Bussen
,
M.
,
Galuppo
,
P.
,
Just
,
S.
,
Rottbauer
,
W.
,
Frantz
,
S.
, et al. 
(
2008
).
MicroRNA-21 contributes to myocardial disease by stimulating MAP kinase signalling in fibroblasts
.
Nature
456
,
980
-
984
.
Tsukazaki
,
T.
,
Chiang
,
T. A.
,
Davison
,
A. F.
,
Attisano
,
L.
and
Wrana
,
J. L.
(
1998
).
SARA, a FYVE domain protein that recruits Smad2 to the TGFβ receptor
.
Cell
95
,
779
-
791
.
van Meeteren
,
L.
and
ten Dijke
,
P.
(
2012
).
Regulation of endothelial cell plasticity by TGF-β
.
Cell Tissue Res.
347
,
177
-
186
.
van Rooij
,
E.
,
Sutherland
,
L. B.
,
Thatcher
,
J. E.
,
DiMaio
,
J. M.
,
Naseem
,
R. H.
,
Marshall
,
W. S.
,
Hill
,
J. A.
and
Olson
,
E. N.
(
2008
).
Dysregulation of microRNAs after myocardial infarction reveals a role of miR-29 in cardiac fibrosis
.
Proc. Natl. Acad. Sci. USA
105
,
13027
-
13032
.
Yao
,
Y.
,
Jumabay
,
M.
,
Ly
,
A.
,
Radparvar
,
M.
,
Cubberly
,
M. R.
and
Bostrom
,
K. I.
(
2013
).
A role for the endothelium in vascular calcification
.
Circ. Res.
113
,
495
-
504
.
Yoshida
,
M.
,
Okubo
,
N.
,
Chosa
,
N.
,
Hasegawa
,
T.
,
Ibi
,
M.
,
Kamo
,
M.
,
Kyakumoto
,
S.
and
Ishisaki
,
A.
(
2012
).
TGF-beta-operated growth inhibition and translineage commitment into smooth muscle cells of periodontal ligament-derived endothelial progenitor cells through Smad- and p38 MAPK-dependent signals
.
Int. J. Biol. Sci.
8
,
1062
-
1074
.
Yoshimatsu
,
Y.
and
Watabe
,
T.
(
2011
).
Roles of TGF-beta signals in endothelial-mesenchymal transition during cardiac fibrosis
.
Int. J. Inflamm.
2011
,
724080
.
Zeisberg
,
E. M.
,
Tarnavski
,
O.
,
Zeisberg
,
M.
,
Dorfman
,
A. L.
,
McMullen
,
J. R.
,
Gustafsson
,
E.
,
Chandraker
,
A.
,
Yuan
,
X.
,
Pu
,
W. T.
,
Roberts
,
A. B.
, et al. 
(
2007
).
Endothelial-to-mesenchymal transition contributes to cardiac fibrosis
.
Nat. Med.
13
,
952
-
961
.
Zeisberg
,
E. M.
,
Potenta
,
S. E.
,
Sugimoto
,
H.
,
Zeisberg
,
M.
and
Kalluri
,
R.
(
2008
).
Fibroblasts in kidney fibrosis emerge via endothelial-to-mesenchymal transition
.
J. Am. Soc. Nephrol.
19
,
2282
-
2287
.
Zhang
,
J.
,
Zhang
,
Z.
,
Zhang
,
D. Y.
,
Zhu
,
J.
,
Zhang
,
T.
and
Wang
,
C.
(
2013
).
microRNA 126 Inhibits the transition of endothelial progenitor cells to mesenchymal cells via the PIK3R2-PI3K/Akt signalling pathway
.
PLoS ONE
8
,
e83294
.
Zhang
,
Z.
,
Zhang
,
T.
,
Zhou
,
Y.
,
Wei
,
X.
,
Zhu
,
J.
,
Zhang
,
J.
and
Wang
,
C.
(
2015
).
Activated phosphatidylinositol 3-kinase/Akt inhibits the transition of endothelial progenitor cells to mesenchymal cells by regulating the forkhead box subgroup O-3a signaling
.
Cell Physiol. Biochem.
35
,
1643
-
1653
.

Competing interests

The authors declare no competing or financial interests.