RhoGEF17—An Essential Regulator of Endothelial Cell Death and Growth

The Rho guanine nucleotide exchange factor RhoGEF17 was described to reside in adherens junctions (AJ) in endothelial cells (EC) and to play a critical role in the regulation of cell adhesion and barrier function. The purpose of this study was to analyze signal cascades and processes occurring subsequent to AJ disruption induced by RhoGEF17 knockdown. Primary human and immortalized rat EC were used to demonstrate that an adenoviral-mediated knockdown of RhoGEF17 resulted in cell rounding and an impairment in spheroid formation due to an enhanced proteasomal degradation of AJ components. In contrast, β-catenin degradation was impaired, which resulted in an induction of the β-catenin-target genes cyclin D1 and survivin. RhoGEF17 depletion additionally inhibited cell adhesion and sheet migration. The RhoGEF17 knockdown prevented the cells with impeded cell–cell and cell–matrix contacts from apoptosis, which was in line with a reduction in pro-caspase 3 expression and an increase in Akt phosphorylation. Nevertheless, the cells were not able to proliferate as a cell cycle block occurred. In summary, we demonstrate that a loss of RhoGEF17 disturbs cell–cell and cell–substrate interaction in EC. Moreover, it prevents the EC from cell death and blocks cell proliferation. Non-canonical β-catenin signaling and Akt activation could be identified as a potential mechanism.


Introduction
Endothelial cells (EC) form a continuous monolayer that covers the luminal side of blood vessels, thereby building a chemicophysical barrier between the blood stream and the interstitium. The integrity of this endothelial barrier is complexly regulated and dependent on cell-cell contacts including adherens junctions (AJ). Besides their function in cell-cell adhesion, AJ play a role in actin cytoskeleton remodeling, intracellular signaling, and transcriptional regulation in EC [1]. The major structural component of AJ in EC is vascular endothelial (VE)-cadherin. In addition, EC express neuronal (N)-cadherin, which was suggested to play a role in stabilizing contacts between EC and mural cells [2][3][4][5][6]. Of note, in addition to primary human umbilical vein EC (HUVEC) immortalized rad fad pad endothelial cells (RFPEC) were used in this study, which do not express significant

RhoGEF17 Is Essential for AJ Formation in EC
In order to investigate the involvement of RhoGEF17 in AJ signaling, we first verified its role in the regulation of AJ integrity. We transduced HUVEC and spontaneously immortalized RFPEC [29], with adenoviruses encoding two shRNAs (sh17-1, sh17-2) targeting RhoGEF17 in its PH domain (Supplementary Materials Figure S1A). For negative control, we used an EGFP-expressing adenovirus (EGFP) and an adenovirus encoding a shRNA against p63RhoGEF (shp63) [27,28], which is not expressed in EC [32]. We show that a downregulation of RhoGEF17 by around 50% in HUVEC and 70% in RFPEC resulted in a gradual cell rounding and loss of cell-cell contacts over time ( Figure 1A,B, Supplementary Materials Figure S1B-D). We substantiated this finding by demonstrating that after the knockdown of RhoGEF17 no RFPEC spheroids were formed ( Figure 1C, Supplementary Materials Figure  S1E). In contrast to what has been shown before, the AJ proteins VE-cadherin and N-cadherin, the dominant isoforms in HUVEC and RFPEC [7], respectively, and p120-catenin were downregulated in sh17-transduced cells compared to both controls ( Figure 1D, Supplementary Materials Figure S1F,G). Taken together this data indicates that a loss of RhoGEF17 in EC results in a loss of AJ proteins and consequently in an impairment of cell-cell interaction.
Cells 2021, 10, x FOR PEER REVIEW 5 of 18 system (Synentech, Elmshorn, Germany). HUVEC were seeded in gelatin-coated 24-well plates and transduced 24 h later. After a further 24, 48, and 72 h bright field and EGFP fluorescence images were taken with a microscope and the cells (EGFPand EGFP + ) were manually counted. Statistics-The statistical analysis was performed using Graph Pad Prism 5 or 7. All results are given as mean+SEM. To test for normal distribution, the F-test was used. Data of two groups were compared using the Student's t-test. The analysis and comparison of three and more groups was performed with the one or two way analysis of variance (ANOVA) and the corresponding post-test. p-values ≤ 0.05 were considered statistically significant and are indicated.

RhoGEF17 is Essential for AJ Formation in EC
In order to investigate the involvement of RhoGEF17 in AJ signaling, we first verified its role in the regulation of AJ integrity. We transduced HUVEC and spontaneously immortalized RFPEC [29], with adenoviruses encoding two shRNAs (sh17-1, sh17-2) targeting RhoGEF17 in its PH domain (Supplementary Materials Figure S1A). For negative control, we used an EGFP-expressing adenovirus (EGFP) and an adenovirus encoding a shRNA against p63RhoGEF (shp63) [27,28], which is not expressed in EC [32]. We show that a downregulation of RhoGEF17 by around 50% in HUVEC and 70% in RFPEC resulted in a gradual cell rounding and loss of cell-cell contacts over time ( Figure 1A,B, Supplementary Materials Figure S1B-D). We substantiated this finding by demonstrating that after the knockdown of RhoGEF17 no RFPEC spheroids were formed ( Figure 1C, Supplementary Materials Figure S1E). In contrast to what has been shown before, the AJ proteins VE-cadherin and N-cadherin, the dominant isoforms in HUVEC and RFPEC [7], respectively, and p120-catenin were down-regulated in sh17-transduced cells compared to both controls ( Figure 1D, Supplementary Materials Figure S1F,G). Taken together this data indicates that a loss of RhoGEF17 in EC results in a loss of AJ proteins and consequently in an impairment of cell-cell interaction.  Values are normalized and given as means + SEM with the single data points, n = 7, * p < 0.05 analyzed by paired t-testing. (B). Depicted are bright field/EGFP overlay images of transduced HUVEC. Scale bar = 100 µm. (C) Rat fat pad endothelial cells (RFPEC) were transduced for 48 h and then used to generate spheroids. Bright field and fluorescence images are shown. Scale bar = 200 µm. (D) VE-cadherin, p120-catenin, and α-tubulin were detected by immunoblot in lysates of transduced HUVEC. Shown are representative immunoblots and the quantified data normalized by α-tubulin and relative to EGFP as means + SEM with the single data points, n = 4 − 7, * p < 0.05 analyzed by paired t-testing.

AJ Proteins Are Predominantly Lost by Proteasomal Degradation in EC with Low Expression of RhoGEF17
Following, we investigated the mechanism behind the down-regulation of AJ proteins as a consequence of the RhoGEF17 knockdown. We used for this purpose RFPEC, which showed comparable results to HUVEC with respect to cell rounding and protein regulation, but could be more efficiently transduced. We demonstrated by a time course that the decline in cadherin protein levels was delayed by around 12 h compared to RhoGEF17 expression ( Figure 2A). To test if this results from a transcriptional change, we performed a qPCR analysis. The knockdown of RhoGEF17 resulted in an expected strong decrease in RhoGEF17 mRNA and a moderate decrease in p120-catenin mRNA. However, neither N-cadherin nor the housekeeping gene PBGD showed any differences in transcript levels ( Figure 2B). Next, we detected N-cadherin and p120-cateinin by immunofluorescence, which demonstrated that both proteins were not only localized at the plasma membrane, but could be additionally found in a punctated pattern inside the cells when RhoGEF17 expression is reduced ( Figure 2C). As these findings suggested that the decline in N-cadherin protein was not based on a transcriptional change, but on a change in protein turnover, we treated the transduced cells with the proteasome inhibitor Bortezomib and performed an immunoblot analysis. Indeed, after inhibition of the proteasome, no differences in the protein content of N-cadherin could be detected in RhoGEF17-knockdown cells compared to control. Likewise, the drop in p120-catenin was prevented by Bortezomib, suggesting that proteasomal degradation and not the decline in gene transcription is the dominant mechanism leading to the p120-catenin downregulation ( Figure 2D). bright field/EGFP overlay images of transduced HUVEC. Scale bar = 100 µm. (C) Rat fat pad endothelial cells (RFPEC) were transduced for 48 h and then used to generate spheroids. Bright field and fluorescence images are shown. Scale bar = 200 µm. (D) VE-cadherin, p120-catenin, and αtubulin were detected by immunoblot in lysates of transduced HUVEC. Shown are representative immunoblots and the quantified data normalized by α-tubulin and relative to EGFP as means + SEM with the single data points, n = 4 − 7, * p < 0.05 analyzed by paired t-testing.

AJ Proteins are Predominantly Lost by Proteasomal Degradation in EC with Low Expression of RhoGEF17
Following, we investigated the mechanism behind the down-regulation of AJ proteins as a consequence of the RhoGEF17 knockdown. We used for this purpose RFPEC, which showed comparable results to HUVEC with respect to cell rounding and protein regulation, but could be more efficiently transduced. We demonstrated by a time course that the decline in cadherin protein levels was delayed by around 12 h compared to RhoGEF17 expression (Figure 2A). To test if this results from a transcriptional change, we performed a qPCR analysis. The knockdown of RhoGEF17 resulted in an expected strong decrease in RhoGEF17 mRNA and a moderate decrease in p120-catenin mRNA. However, neither N-cadherin nor the housekeeping gene PBGD showed any differences in transcript levels ( Figure 2B). Next, we detected N-cadherin and p120-cateinin by immunofluorescence, which demonstrated that both proteins were not only localized at the plasma membrane, but could be additionally found in a punctated pattern inside the cells when RhoGEF17 expression is reduced ( Figure 2C). As these findings suggested that the decline in N-cadherin protein was not based on a transcriptional change, but on a change in protein turnover, we treated the transduced cells with the proteasome inhibitor Bortezomib and performed an immunoblot analysis. Indeed, after inhibition of the proteasome, no differences in the protein content of N-cadherin could be detected in RhoGEF17-knockdown cells compared to control. Likewise, the drop in p120-catenin was prevented by Bortezomib, suggesting that proteasomal degradation and not the decline in gene transcription is the dominant mechanism leading to the p120-catenin downregulation ( Figure  2D). Non-transduced (nt) cells were used as additional control. RhoGEF17, N-cadherin and p120-catenin were detected by immunoblot in whole cell lysates. Shown are representative immunoblots of RhoGEF17, N-cadherin, p120-catenin and α-tubulin (left) and the quantitative analyses. Values were normalized by α-tubulin and are given relative to non-transduced cells treated with DMSO only. Shown are means + SEM and the single data points; n = 3-5, * p < 0.05 assessed by 2-way ANOVA with Tukey's multiple comparison testing.

The Reduction of RhoGEF17 Is Accompanied by an Increased Protein Modification of β-Catenin and Changes in the Expression of Associated Genes
AJ dissociation results in non-canonical Wnt signaling by the release of the major Wnt pathway mediator β-catenin [33]. Therefore, we studied the impact of the RhoGEF17 knockdown on β-catenin by immunoblot analysis. We detected two distinct β-catenin signals of which the signal with the higher molecular weight, named as modified β-catenin (mod. β-catenin), was increased after depletion of RhoGEF17 in HUVEC and RFPEC ( Figure 3A,B, Supplementary Materials Figure S2A). The total β-catenin content (intensity of both signals together) was, however, unchanged ( Figure 3A,B). In order to characterize the nature of the two β-catenin variants, we treated RFPEC with Bortezomib. Inhibition of the proteasome had no impact on total β-catenin levels, but shifted the ratio of both variants in favor of the larger one. We further used a phospho-specific anti-β-catenin antibody recognizing its phosphorylation at serine 33, serine 37 and threonine 41. This allowed us to demonstrate that inhibition of the proteasome, and more importantly RhoGEF17 depletion, resulted in an accumulation of phosphorylated β-catenin ( Figure 3C). We next analyzed the impact of the RhoGEF17 knockdown on the expression of the inhibitory Wnt pathway component axin1 and the two downstream targets survivin and cyclin D1. Surprisingly, we found all three proteins to be increased after the knockdown of RhoGEF17 ( Figure 3D). We further verified for axin1 and cyclin D1 that this increase was due to an enhanced transcription (Supplementary Materials Figure S2B). As this implied that despite the increase in β-catenin phosphorylation and the higher expression of the Wnt inhibitor axin1, a net activation of the Wnt pathway had occurred, we investigated the distribution of β-catenin in the cytosolic and nuclear fractions of the transduced RFPEC. In line, we found more unmodified, transcriptional active β-catenin in the nuclear fraction of RhoGEF17 knockdown cells compared to control transduced cells, whereas in the cytosolic fraction the opposite distribution was detectable ( Figure 3E).

The Reduction of RhoGEF17 Impairs RFPEC Adhesion and Sheet Migration
Based on the observed cell rounding of RhoGEF17 knockdown cells, we analyzed next the adhesion behavior of transduced RFEPC and could show that the knockdown of RhoGEF17 strongly impaired the adhesion ability ( Figure 4A, Supplementary Materials Figure S3A) and maximal cell spreading reflected by the reduced cell area ( Figure 4B, Supplementary Materials Figure S3B). By staining of the focal adhesion protein vinculin, we were able to demonstrate that the knockdown of RhoGEF17 resulted in a loss of central focal adhesion sites in transduced (EGFP + ) cells, but also in non-transduced (EGFP -) adjacent cells ( Figure 4C,D). The expression level of vinculin was unaltered ( Figure 4E). We could further show that the loss in focal adhesion sites was accompanied by a loss in prominent actin fibers ( Figure 4F). As cell adhesion and migration are strongly connected processes, we studied next the influence of the RhoGEF17 knockdown on sheet migration. The work presented by Mitin and colleagues in 2013 already demonstrated that a knockdown of RhoGEF17 impairs the migration directionality of EC in a single cell migration assay [24]. We found that the knockdown of RhoGEF17 fully impaired the coordinated closure of the cell-free area in a scratch assay. Interestingly, more single cells were present in the "wound"

The Reduction of RhoGEF17 Impairs RFPEC Adhesion and Sheet Migration
Based on the observed cell rounding of RhoGEF17 knockdown cells, we analyzed next the adhesion behavior of transduced RFEPC and could show that the knockdown of RhoGEF17 strongly impaired the adhesion ability ( Figure 4A, Supplementary Materials Figure S3A) and maximal cell spreading reflected by the reduced cell area ( Figure 4B, Supplementary Materials Figure S3B). By staining of the focal adhesion protein vinculin, we were able to demonstrate that the knockdown of RhoGEF17 resulted in a loss of central prominent actin fibers ( Figure 4F). As cell adhesion and migration are strongly connected processes, we studied next the influence of the RhoGEF17 knockdown on sheet migration. The work presented by Mitin and colleagues in 2013 already demonstrated that a knockdown of RhoGEF17 impairs the migration directionality of EC in a single cell migration assay [24]. We found that the knockdown of RhoGEF17 fully impaired the coordinated closure of the cell-free area in a scratch assay. Interestingly, more single cells were present in the "wound" (Figure 4G, Supplementary Materials Figure S3C).

The Reduction of RhoGEF17 Prevents Apoptosis and Induces a Cell Cycle Block in EC
Adherent cells, like EC, which experience disturbances in their cell-cell and cellmatrix contacts undergo anchorage-dependent apoptosis, also termed anoikis [34]. We wondered if the knockdown of RhoGEF17 and thus the disruption of AJ and the weakening of focal adhesions consequently result in a higher apoptosis rate. We first performed an annexin V staining of semi-efficiently transduced cells and counted EGFP + /annexin V + and EGFP -/annexin V + cells. We show that the percentages of apoptotic transduced HUVEC and RFPEC did not differ between the conditions, but for the non-transduced cells significant higher percentages of apoptotic cells were found in the conditions were the sh17 viruses were used. In general, the basal apoptosis rate of HUVEC was higher as of the immortalized RFPEC ( Figure 5A,B, Supplementary Materials Figure S4A). This data indicated that a knockdown of RhoGEF17 does not only impair cell-cell and cell-matrix contacts, but surprisingly also protect the rounded knockdown cells from apoptosis. To further validate this unexpected finding, we used 100% transduced RFPEC and measured the expression of pro-caspase 3 and cleaved caspase 3 in lysates by immunoblot. In line with the annexin V staining, we found a lower expression of pro-caspase 3 and fewer cleaved caspase 3 in RFPEC with lowered RhoGEF17 expression compared to controls ( Figure 5C, Supplementary Materials Figure S4B). Moreover, the pro-survival kinase Akt displayed a higher degree of phosphorylation in these cells ( Figure 5D). We wondered whether the reduction in RhoGEF17 influences the cell cycle activity and proliferation capacity of the transduced cells. Therefore, cell cycle progression was investigated by flow cytometry experiments with propidium iodide (PI)-stained EC. The knockdown of RhoGEF17 in HUVEC resulted in a higher proportion of cells in G2/M phase ( Figure 5E, Supplementary Materials Figure S4C) and in RFPEC more cells were found to be in S-and G2/M-phase ( Figure 5F, Supplementary Materials Figure S4D). Finally, we were able to demonstrate that the down-regulation of RhoGEF17 in EC inhibited cell proliferation. For RFPEC we show that an efficient RhoGEF17 knockdown inhibits proliferation compared to EGFPtransduced cells ( Figure 5G). Importantly, we transduced HUVEC with varying amounts of the sh17-1 adenovirus to obtain low, medium, and high transduction efficiencies (12.6%, 53.6%, and 74.5%). We found that the number of EGFP + cells stayed constant over two days independent of the transduction efficiency, thus demonstrating that no proliferation of cells with reduced RhoGEF17 expression occurred. We further counted the EGFPcells in these experiments, which demonstrated that in the presence of only a minor portion of EGFP + cells (Low), the EGFPcells proliferated similar to the shp63 control. However, in the presence of higher numbers of EGFP + cells, the proliferation of EGFPcells was blunted for medium transduction efficiency conditions and the number of non-transduced cells even decreased when the RhoGEF17 knockdown cells constituted the majority of cells ( Figure 5H, Supplementary Materials Figure S4E). Therefore, this data supports not only the RFPEC proliferation data, but also our apoptosis data, demonstrating that RhoGEF17 knockdown cells were protected from apoptosis, whereas the by neighboring RhoGEF17 knockdown cells isolated wildtype cells underwent apoptosis.

Discussion
RhoGEF17 was described as a RhoA, B, and C-specific GEF [22], which was first identified in 2002 as a member of the Dbl family of RhoGEFs [23]. In EC, RhoGEF17 was found amongst the highest expressed RhoGEFs and implicated to play a role in cancerassociated angiogenesis [20,35]. In HUVEC, it was shown that a knockdown of RhoGEF17 impaired AJ integrity, the formation of tube-like structures in an in vitro angiogenesis assay, and the persistence of single cell migration [19,24].
With our study we confirmed the role of RhoGEF17 in the regulation of AJ and in cell migration, however, the strength of the phenotype was more pronounced compared to the previously published studies [19,24]. We have not only seen an impairment in AJ function, but we found that the knockdown of RhoGEF17 led to a loss of cell-cell interaction in 2D and 3D cultures. On a molecular level, we could prove that AJ proteins were downregulated by proteasomal degradation. This included N-cadherin in RFPEC, which do not express VE-cadherin and were shown to use instead N-cadherin for EC-EC interaction [7], and VE-cadherin in HUVEC as well as p120-catenin in both cell types. Although, HUVEC do express both VE-and N-cadherin, we did not investigate changes in N-cadherin, as this cadherin isoform was suggested to play a role in heterocellular contacts, like between EC and mural cells [3][4][5][6].
Furthermore, in our experiments we found a strongly impaired cell adhesion capacity and the RhoGEF17 knockdown cells were completely unable to migrate in a sheet. As a possible explanation for the differences in phenotype severity, the methods of transduction could be accounted. Both Ngok and Mitin and coworkers used a lentiviral approach, whereas we used an adenoviral approach. In fact, it has been shown that adenoviruses are more efficient in gene transfer for EC than lentiviruses, suggesting that the achieved knockdown in our study was more pronounced and likely faster in onset [36,37].
Beyond RhoGEF17 s role in the regulation of AJ integrity, cell adhesion, and migration, we describe in our study its role in cell death, cell cycle control, and proliferation. The survival and growth of EC is strongly dependent on cell-cell and cell-matrix contacts. With respect to cell-cell contacts, it was shown that the deficiency of VE-cadherin, its cytosolic truncation, and the interference of its interaction by monoclonal antibodies results in endothelial cell apoptosis [38,39]. A loss in EC-matrix contacts also induced apoptosis, as demonstrated by exposing EC to non-adhesive surfaces, or by using antagonizing antibodies against integrins and RGD mimics [40][41][42][43]. In line with these observations, we found that EC, treated with, but not transduced by the sh17 adenovirus, showed a higher apoptosis rate as EC with cell-cell and cell-matrix contacts under control conditions. However, rounded EC with a reduction in RhoGEF17 expression did not demonstrate an increased apoptosis rate, suggesting that RhoGEF17 is involved in anoikis in EC. On a molecular level we demonstrated that RhoGEF17 knockdown cells expressed more survivin, less pro-and cleaved caspase 3, and showed an increase in Akt phosphorylation, which reflects together an anti-apoptotic phenotype. These changes are typical for anoikisresistent cells, like metastatic cancer cells [44,45]. Also, for anoikis-resistent EC it has been described that an increase in Akt phosphorylation is essential for cell survival [34,46]. However, in contrast to anoikis-resistent EC, which are able to proliferate, RhoGEF17depleted EC showed no proliferation capacity at all. Cell cycle analysis revealed that the fraction of RFPEC and HUVEC in G1 phase were strongly reduced and instead the percentages of cells in S-and/or G2/M-phase were increased. This data argues for a block in cell cycle, which consequently impairs proliferation. In accordance to the lower number of cells in the G1 phase, we could show that the cell cycle regulator cyclin D1 was increased. The induction of cyclin D1, and its binding to CDK4 or CDK6, was demonstrated to be a rate-limiting event during cell-cycle progression through G1 phase [47,48]. However, the exact mechanisms behind the arrest is not clear. As an explanation, a blockade of the proteasome, due to the sudden increase in substrate load by AJ disruption, could serve. For HUVEC it was demonstrated that proteasome inhibition with Bortezomib resulted in G2/M arrest and subsequently in cell death [49]. In addition, RhoGEF17 was described to be part of the spindle assembly checkpoint complex. Its depletion in HeLa cells resulted in an acceleration of mitosis, which impaired chromosome congression, biorientation, and segregation [50]. It could therefore be speculated that this novel role of RhoGEF17 finally induces a blockade of the cell cycle.
A possible signaling hub between RhoGEF17 and the described functional changes could be the multifunctional protein β-catenin. When RhoGEF17 is less expressed and cadherins are degraded, the bound junctional β-catenin is released and becomes part of the β-catenin signaling pool. This led in our study to an up-regulation of the two β-catenin target genes cyclin D1 and survivin [51,52]. Interestingly, in our immunoblot studies we detected two β-catenin bands of which we named the variant with the higher molecular weight "modified" β-catenin. Under control conditions the ratio of this modified variant compared to the variant with the lower molecular weight was around one to four and in RhoGEF17 knockdown cells the ratio was changed to one to two. When the cells were treated with Bortezomib the amount of modified β-catenin increased in the controls and was then similar under all conditions. This implies that in RhoGEF17 knockdown cells an accumulation of β-catenin takes place which can escape from proteasomal degradation. We further show that the modified β-catenin can be detected by a phospho-β-catenin antibody. Whether or not the RhoGEF17-dependent shift in β-catenin size results from phosphorylation and/or other modifications is not clear [53][54][55]. At least the shift size of around 20 kDa is indicative that larger modifications, like ubiquitination, are contributing to it [56,57]. Therefore, our data indicate that a RhoGEF17 knockdown in EC results in an accumulation of posttranslationally modified β-catenin ( Figure 6). The most likely explanation for this is an overload of the proteasome by the massive degradation of other destabilized AJ proteins. In general, an increase in the cytosolic pool of β-catenin is thought to be associated with enhanced nuclear translocation and thus enhanced transcription of β-catenin dependent genes. In line with this interpretation, we detected an increase in the nuclear pool of the unmodified β-catenin as well as an upregulation of β-catenin dependent gene products like cyclin D1 and survivin after RhoGEF17 depletion. In this context, the increased activity of Akt in RhoGEF17 knockdown cells might be of importance as it phosphorylates β-catenin at serine 552 ( Figure 6). This leads to β-catenin stabilization via binding to the scaffold protein 14-3-3ζ. It was further hypothesized that this phosphorylation induces a conformational change disengaging the binding of β-catenin from the destruction complex and resulting in an increased transcriptional activity, independent of GSK3β-mediated phosphorylation [58,59].
In conclusion, we show that RhoGEF17 is not only an integral part of the AJ complex in EC and thus contributes to the maintenance of cell-cell contacts and to cell migration, but has also a so far unknown role in the regulation of cell death and growth. Whether this mechanism applies only for conditions resulting from an accidental loss of RhoGEF17, as rarely found in patients with intracranial aneurysms, or exists as a phenotypic state, remains to be investigated. Based on the cellular phenotype, RhoGEF17 downregulation could play a role in, for example, EC senescence or as an intermediate step in endothelial to mesenchymal transition (EndMT). Typical features of senescent EC, which resembles the phenotype of RhoGEF17 knockdown EC, are a reduction in AJ integrity, higher expression of cyclin D1, an enhanced phosphorylation of Akt, and lowered proliferation and migration capacities. However, the cell cycle block in senescent EC occurs in G1 phase and the cells possess a larger cell surface, which contradicts our findings [60][61][62][63]. In EndMT, a loss in AJ proteins is a common feature and the signaling involves β-catenin and Akt phosphorylation. Moreover, it is likely that during EndMT the cells lose their ability to migrate in a sheet, but instead migrate as single cells as seen for RhoGEF17 knockdown cell [64,65]. Targeted approaches are necessary to gain deeper insight in RhoGEF17 s regulation and function in vivo. migration capacities. However, the cell cycle block in senescent EC occurs in G1 phase and the cells possess a larger cell surface, which contradicts our findings [60][61][62][63]. In EndMT, a loss in AJ proteins is a common feature and the signaling involves β-catenin and Akt phosphorylation. Moreover, it is likely that during EndMT the cells lose their ability to migrate in a sheet, but instead migrate as single cells as seen for RhoGEF17 knockdown cell [64,65]. Targeted approaches are necessary to gain deeper insight in RhoGEF17′s regulation and function in vivo. Figure 6. Scheme of RhoGEF17 function in EC. RhoGEF17 stabilizes AJ in EC. Its loss leads to AJ protein degradation via the proteasome and an accumulation of β-catenin in its destruction complex (phosphorylated and ubiquitinated) form. A part of the β-catenin pool, which might be phosphorylated by Akt, is translocated to the nucleus and induces β-catenin/TCF dependent gene transcription of, for example, survivin and cyclin D1. Besides impaired cell adhesion and migration due to the disruption of the AJ, RhoGEF17-depleted EC can escape anoikis and end in cell cycle arrest. In contrast neighboring cells, without RhoGEF17 knockdown, enter apoptosis due to the loss of cell-cell contacts. EC = Endothelial cell, RhoGEF17 = Rho-specific guanine nucleotide exchange factor 17, p120 = p120-catenin, α = α-catenin, TCF = Transcription factor, Ub = ubiquitinated, P = phosphorylated.

Supplementary Materials:
The following are available online at www.mdpi.com/xxx/s1, Figure S1: RhoGEF17 is essential for cell-cell contacts and AJ protein regulation in EC; Figure S2:RhoGEF17 regulates β-catenin; Figure S3: The shp63 adenovirus has no effect on RFPEC adhesion, cell size, and sheet migration; Figure S4: The shp63 adenovirus has no effect on EC apoptosis, caspase 3 expression, and cell cycle. Figure 6. Scheme of RhoGEF17 function in EC. RhoGEF17 stabilizes AJ in EC. Its loss leads to AJ protein degradation via the proteasome and an accumulation of β-catenin in its destruction complex (phosphorylated and ubiquitinated) form. A part of the β-catenin pool, which might be phosphorylated by Akt, is translocated to the nucleus and induces β-catenin/TCF dependent gene transcription of, for example, survivin and cyclin D1. Besides impaired cell adhesion and migration due to the disruption of the AJ, RhoGEF17-depleted EC can escape anoikis and end in cell cycle arrest. In contrast neighboring cells, without RhoGEF17 knockdown, enter apoptosis due to the loss of cell-cell contacts. EC = Endothelial cell, RhoGEF17 = Rho-specific guanine nucleotide exchange factor 17, p120 = p120-catenin, α = α-catenin, TCF = Transcription factor, Ub = ubiquitinated, P = phosphorylated.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/cells10040741/s1, Figure S1: RhoGEF17 is essential for cell-cell contacts and AJ protein regulation in EC; Figure S2:RhoGEF17 regulates β-catenin; Figure S3: The shp63 adenovirus has no effect on RFPEC adhesion, cell size, and sheet migration; Figure S4: The shp63 adenovirus has no effect on EC apoptosis, caspase 3 expression, and cell cycle.

Acknowledgments:
We thank Kristina Stephan-Schnatz, Heike Rauscher and Beate Ramba for their technical support in this project. Furthermore, we want to thank Karin Bieback and Stefanie Uhlig from the Cell Sorting Core Facility FlowCore Mannheim for their help with the flow cytometry measurements and Cleo-Aron Weis from the Institute of Pathology Mannheim for sharing the survivin antibody with us.

Conflicts of Interest:
The authors declare no conflict of interest.

AJ
Adherens junctions EC Endothelial cells FA focal adhesion HUVEC Human umbilical vein endothelial cells RFPEC Rat fat pad endothelial cells RhoGEF Rho-specific guanine nucleotide exchange factor TEM4 Tumor endothelial marker 4