Phorbol esters induce PLVAP expression via VEGF and additional secreted molecules in MEK1‐dependent and p38, JNK and PI3K/Akt‐independent manner

Abstract Endothelial diaphragms are subcellular structures critical for mammalian survival with poorly understood biogenesis. Plasmalemma vesicle associated protein (PLVAP) is the only known diaphragm component and is necessary for diaphragm formation. Very little is known about PLVAP regulation. Phorbol esters (PMA) are known to induce de novo PLVAP expression and diaphragm formation. We show that this induction relies on the de novo production of soluble factors that will act in an autocrine manner to induce PLVAP transcription and protein expression. We identified vascular endothelial growth factor‐A (VEGF‐A) signalling through VEGFR2 as a necessary but not sufficient downstream event as VEGF‐A inhibition with antibodies and siRNA or pharmacological inhibition of VEGFR2 only partially inhibit PLVAP upregulation. In terms of downstream pathways, inhibition of MEK1/Erk1/2 MAP kinase blocked PLVAP upregulation, whereas inhibition of p38 and JNK MAP kinases or PI3K and Akt had no effect on PMA‐induced PLVAP expression. In conclusion, we show that VEGF‐A along with other secreted proteins act synergistically to up‐regulate PLVAP in MEK1/Erk1/2 dependent manner, bringing us one step further into understanding the genesis of the essential structures that are endothelial diaphragms.

capillaries results in protein losing enteropathy, hypoproteinemia and hypertriglyceridemia causing a kwashiorkor-like wasting syndrome and death. 3,6 Interestingly, PLVAP reconstitution in the endothelial compartment in mice restores diaphragms exclusively in EC in vascular beds where the diaphragms are native, demonstrating that additional factors are required for diaphragm formation. 3 Very little is known on PLVAP and diaphragm regulation despite their importance. Phorbol esters such as phorbol myristate acetate (PMA) are known to induce robust de novo formation of fenestrae and transendothelial channels with their associated diaphragms in primary EC in culture. 18 PMA also induces diaphragms of caveolae and PLVAP expression in MEK1-dependent and PKC-independent manner. 16 Of note, PMA is also a known secretagogue in human EC. 19 Vascular endothelial growth factor-A (VEGF-A) was also shown to be essential for formation and maintenance of fenestrae with diaphragms. VEGF-A (and not FGF-2 or VEGF-C) induces new vessels with fenestrae with diaphragms [20][21][22][23][24][25] in Rac1 dependent manner. 22 Deletion of VEGF-A in the kidney podocytes, pancreas epithelial cells or hepatocytes [26][27][28] or systemically delivered VEGFR2 inhibitors 29 result in loss of fenestrae in mice. While overwhelmingly clear in the case of fenestrae, the effect of VEGF-A/VEGFR2 signalling on PLVAP expression modulation seems to be context dependent.
While a VEGFR2 receptor-selective engineered form of VEGF-A upregulates PLVAP expression in single donor human umbilical vein EC (HUVEC), 30 primary EC cultured in presence of VEGF-A do not or poorly express PLVAP 16 and VEGF-A has no effect 17 or even decreases 31 PLVAP expression in immortalized mouse EC lines that constitutively express PLVAP. Moreover, VEGFR2 signalling inhibition in vivo does not modify PLVAP expression in the lung. 32 Finally, the downregulation of PLVAP in specialized vascular beds forming the blood-brain barrier, [33][34][35][36] or in developing arteries, glomeruli and cell culture [37][38][39][40] appear to be controlled by the Wnt and Notch signalling pathways respectively.
In order to arrive at a cell culture system where fenestrae with diaphragms and PLVAP could be induced at high frequency by physiological cues in primary EC, we have sought to dissect the molecular mechanism of PLVAP upregulation by PMA. We show that PMA upregulation of PLVAP mRNA and protein depends on de novo protein synthesis and secretion of a group of proteins that act synergistically in autocrine fashion. Among these secreted proteins, we identified VEGF-A signalling through VEGFR2 as important but not sufficient for PLVAP expression. In addition, we show that PLVAP upregulation by the PMA-induced secreted factors is MEK1-dependent and JNK-, p38-, PI3K-and Akt-independent.
points, supernatant and cells were harvested and further processed for protein or RNA analysis.

| Protein synthesis inhibition with cycloheximide
For chronic PMA and for conditioned medium (CM) treatments, EC were seeded in duplicate on gelatin-coated plates, grown to near confluence, serum starved 1.5 hours in EBM2 and 30 minutes presence of 10 μg/mL CHX in EBM-BSA and stimulated for the duration of the experiment with 50 nmol/L PMA + 10 μg/mL CHX or with 4-6 hours CM + 10 μg/mL CHX. For pulsed PMA treatment, the difference was that EC were stimulated with PMA/CHX for only 30 minutes followed by chase in EBM-FBS containing 10 μg/mL CHX. At indicated time points, cells were rinsed twice in DPBS and lysed for RNA or protein analysis.

| Heparin depletion of conditioned medium
Conditioned medium peaks (4-6 and 6-8 hours) were collected from donor cells cultured in six well plates. For each peak the respective CM was pooled in a 15 mL tube and split into two halves. One half was left untreated (control), the other half of the volume was added to 1 mL settled gel of heparin-agarose previously equilibrated (3×, 5 minutes, RT) in EBM-BSA. The mixture was further incubated (1 hour, RT) with gentle end-over-end rotation before the beads were pelleted by centrifugation (600 g, 10 minutes, RT). Two mL per well control CM or heparin-depleted

| Heat inactivation of CM
Conditioned medium "peaks" were collected after PMA treatment and split into two equal volumes: one half was heat inactivated (45 minutes, 60°C followed by 2 minutes on ice) and the other one left untreated (control) before transfer to serum starved acceptor cells.

| Pertussis toxin treatment
Acceptor cells were serum starved and then treated for 24 hours with CM in presence or absence of 0.1 μg/mL pertussis toxin (PT) (Sigma, cat# P7208).

| Statistics
Data were analysed using Student's t test. P < 0.05 was taken as the level of significance.

| Upregulation of PLVAP mRNA by PMA requires protein translation
In a first step, we asked whether PMA-induced PLVAP mRNA transcription depended on de novo protein synthesis. To answer this, we treated primary human HDMVECn with 50 nmol/L PMA (concentration demonstrated to up-regulate PLVAP and induce the formation of endothelial diaphragms and fenestrae 16 ) in presence or absence of CHX, a protein synthesis inhibitor. 44 As shown previously, 16 cells were exposed to PMA for the entire duration of the experiment.
PLVAP ****mRNA significantly increased in time-dependent manner starting at~2 hours after PMA treatment onset ( Figure 1A). However, there was no increase of PLVAP mRNA or protein ( Figure 1B) when cells were treated with PMA in presence CHX for up to 8 hours of treatment, demonstrating that PLVAP upregulation by PMA requires de novo protein synthesis. To gain insight into the chemical nature of the PLVAP-inducing soluble factor(s), we depleted CM peaks (4-6 and 6-8 hours CM) of heparin-binding proteins. As shown in ( Figure 4B, left), the depletion led to marked decreased in CM ability to induce PLVAP protein in naive acceptor cells ( Figure 4B, right). Interestingly, the residual activity could not be eliminated even after passages over two sequential heparin columns, (data not shown).
Endothelial cells produce chemokines, secreted factors that can bind heparin. 45 We therefore tested the ability of 4-6 and 6-8 hours CM peaks to up-regulate PLVAP in presence of PT, a general/broad  Figure 4C, left). Treatment of acceptor EC with PT had no effect on PLVAP protein upregulation by the 4-6 and 6-8 hours CM peaks ( Figure 4C, right), ruling out a role for chemokine signalling in this system.

| PMA up-regulates PLVAP in part via VEGF/ VEGFR2 signalling
Phorbol myristate acetate up-regulates VEGF-A a known heparinbinding growth factor 48 and its receptors, VEGFR1 and VEGFR2 in HUVEC and HDMVEC, 49 making it them as good candidates for PLVAP upregulation by PMA in EC.
As seen in Figure 5A Figure 5E). However, treatment of EC with up to 40 ng/ml VEGF in addition to PMA does not further increase PLVAP protein levels ( Figure 5F), suggesting that VEGF-A acts downstream of PMA.
To determine which VEGFR is required for PMA/CM mediated PLVAP upregulation pharmacologic inhibitors with different selectivity for VEGFR1, 2 and 3 ( Table 1)   However, while our data do not support a role for p38 signalling in PMA or CM induced PLVAP upregulation, these results are puzzling given the role of VEGF/VEGFR2 signalling in this process.
Others have suggested that VEGF-A regulates PLVAP expression in PI3 kinase-dependent manner. 30 HDMVEC express all four isoforms of p110 (α/PIK3CA, β/PIK3CB, γ/PIK3CG and δ/PIK3CD) and the respective p85/p55/p150 regulatory subunits (PIK3R1-4). Using novel and more selective PI3K pharmacological inhibitors that are in clinical trials such as pictisilib (highly selective for PI3Kα/δ and 11-25 fold lower selectivity on PI3Kβ/γ) and idelalisib (selective for PI3Kγ>δ), we show that PI3K inhibition does not inhibit PLVAP upregulation by PMA or CM. Accordingly, the inhibition of Akt1-3, a major downstream target of PI3K, does not impact PLVAP upregulation. To note, wortmannin, a pan PI3K inhibitor, is partially effective at the larger concentration of 10 μmol/L either suggesting off-target effects or a role for p110β, which is not covered as well by the other inhibitors used. However, the latter is less likely as the dose of pictisilib we used is several orders of magnitude larger than the IC50 for p110β.
In summary, we find that PLVAP upregulation by PMA requires de novo synthesis of multiple secreted proteins that act in an autocrine manner. One of the soluble factors involved is VEGF-A acting through VEGFR2. The signalling is dependent on MEK1/ERK1/2 and independent of p38, JNK, PI3K and Akt1-3. Further transcriptomic and proteomic studies should identify the factors (or combinations thereof) contributing to PLVAP regulation.