S1P-Mediated Endothelial Barrier Enhancement: Role of Rho Family GTPases and Local Lamellipodia

The endothelial cells lining the inner surface of the tissue capillaries and post-capillary venules form a semi-permeable barrier between the blood circulation and interstitial compartments. The semi-permeable barrier in these vessels is the major site of blood-tissue exchange. A compromised endothelial barrier contributes to the pathological process such as edema, acute respiratory distress syndrome (ARDS) and tumor metastasis. Sphingosine-1-phosphate (S1P), an endogenous, bioactive lipid present in all cells, is a potential therapeutic agent that can restore compromised endothelial barrier function. On the other hand, S1P also has pleotropic effects and can either increase or decrease arterial tone and tissue perfusion under different conditions. The detailed mechanisms underlining S1P’s endothelial barrier protective effect are still largely unknown, but are suggested to depend on cell spreading termed “lamellipodia”. Therefore, to fully take advantage of the beneficial properties of S1P, it is important to first understand how S1P-induced lamellipodia protrusions correlate with its effect on endothelial barrier function. It is also important to know the underlining mechanisms that S1P enhances endothelial barrier function, including intracellular signaling and receptor signaling. To study local lamellipodia activities, we acquired time-lapse images of live endothelial cells expressing GFP-actin, and subsequently analyzed different lamellipodia parameters. Experiments were performed under baseline conditions, and during endothelial barrier disruption or enhancement. The compounds used in these experiments included thrombin and S1P. Transendothelial electrical resistance (TER) served as an index of endothelial barrier function for in vitro studies. Changes of local lamellipodia dynamics and endothelial barrier function within the same time frame were studied. For mechanistic studies, we combined biochemical, immunological and pharmacological approaches. Rho family small GTPase activities were measured with an ELISA pull-down assay. Fluorescence Resonance Energy Transfer (FRET) was also used to study the localization of RhoA activation. Pharmacological compounds targeting intracellular signaling messengers were used to test the involvement of Rac1, RhoA, MLC-2 in endothelial local lamellipodia activity and S1P-mediated endothelial barrier enhancement. Receptor agonists and antagonists were used to study the involvement of S1P receptor signaling. Finally, for cell behavior and cytoskeleton studies, we utilized immunofluorescence labeling that enables direct visualization of changes in cytoskeleton, cell-cell junction and focal adhesions. We found that S1P increases both local lamellipodia protrusions and TER. The rapid increase in local lamellipodia protrusion frequencies also corresponded to the rapid increase in TER seen within the same time frame. Under the microscope, local lamellipodia protrusions from adjacent cells overlapped with each other and extended beyond junctional cell-cell contacts. Strikingly, S1P-induced lamellipodia protrusions carry VE-cadherin molecules to the cell-cell contact, established junctional adhesions. Combined with our previous published studies on thrombin induced lamellipodia activity changes, we think lamellipodia protrusions are a major component that regulates endothelial barrier function. Combined, our imaging studies revealed the mechanisms on how lamellipodia regulates endothelial barrier function: 1) lamellipodia overlap and increase the apical to basal diffusion distance, which in turn decreases permeability and upregulates endothelial barrier function. 2) Local lamellipodia protrusions contain VE-cadherin, which is delivered the to the cell-cell contact by the lamellipodia to increase junctional stability. S1P is effective for rescuing thrombin-induced endothelial barrier dysfunction. The known barrier disruptor thrombin, decreased local lamellipodia protrusions, disrupted VE-cadherin integrity, and caused a drop in TER. S1P increased local lamellipodia protrusions after thrombin challenge, and resulted in faster recovery towards baseline TER compared with vehicle controls. Interestingly, we also found that both thrombin and S1P increased MLC-2 phosphorylation at Thr18/Ser19. We subsequently accessed Rho family GTPase activity after thrombin and S1P. As expected, thrombin rapidly increased GTP-bound RhoA levels, and decreased GTP-bound Rac1 levels. Unexpectedly, S1P not only increased GTP-bound Rac1, but also increased GTP-bound RhoA to a more prominently levels (4-fold). Since Rac1 has been implicated in promoting lamellipodia protrusions, we tested the role of Rac1 on the local lamellipodia activities first. We found that Wild-Type (WT) Rac1 group had the highest local lamellipodia protrusion frequencies, protrusion distances, withdraw time and highest percentage of protrusions that lasted more than 5 min. WT Rac1 overexpression had greatest protrusion frequencies and lowest monolayer permeability to FITC-albumin compared to GFP and DN-Rac1 overexpression monolayers. These results suggest that Rac1 is important for baseline endothelial barrier function. This is also confirmed by the finding that pharmacological inhibition of Rac1 significantly decreased baseline TER. Although Rac1 is important for baseline endothelial barrier function, we noticed that it is dispensable in S1P-mediated endothelial barrier enhancement. Rac1 inhibitors, DN-Rac1 overexpression, and Rac1 siRNA knockdown all failed to abolish the S1P-mediated increase in TER. This is partially explained by the findings that S1P-induced Rac1 activation is short-lived and less pronounced in contrast to RhoA activation. We subsequently tested the role of RhoA in S1P-mediated endothelial barrier enhancement, based on our findings that both S1P and thrombin significantly activated RhoA and induced MLC-2 phosphorylation. Significant RhoA activation was found to be mainly at cell periphery and lamellipodia protrusions in HUVEC on FRET, after S1P was given. In addition, RhoA inhibitors significantly decreased the amplitude of S1P-induced MLC-2 phosphorylation, vinculin redistribution and barrier enhancement. The data suggest that the mechanisms involved in S1P-mediated endothelial barrier enhancement depend on RhoA activation and subsequent cytoskeletal rearrangement. We next investigated which receptor is responsible for the endothelial barrier enhancement of S1P. However, antagonism of S1P1, S1P2 or S1P3 alone with W146, JTE-013 or TY-52156 respectively all failed to attenuate S1P-mediated increase in TER. While agonism of S1P1 with CYM-5442 hydrochloride alone produced significant increase in TER, neither S1P2 nor S1P3 activation (CYM 5520 & CYM 5541) produced any change on TER. Interestingly, S1P1 antagonist failed to block the effect of S1P1 agonist on TER. This could be due to that the S1P1 agonist may not be very selective at concentrations tested. We also identified that S1P4 and S1P5 are present on endothelial cells. Further studies would be necessary to elucidate the roles of newly identified S1P4 or S1P5 alone on endothelial barrier function. It is also worth investigating in the future if multiple S1P receptors are involved in its endothelial barrier enhancing effect. In conclusion, we found that lamellipodia protrusions contribute to the endothelial barrier enhancement of S1P. While Rac1 is important for the maintenance of endothelial barrier function, it is dispensable in S1P-mediated endothelial barrier enhancement. On the other hand, RhoA activation appears to be, at least in part, responsible for the endothelial barrier enhancement of S1P. It is currently still unclear if S1P’s endothelial barrier enhancing effect is through one single receptor activation or activation of multiple receptors. Future studies are needed to elucidate the receptor signaling that contributes to S1P-mediated endothelial barrier enhancement