Effect of Lipoxin A4 on Lipopolysaccharide-Induced Endothelial Hyperpermeability

Excessive oxidative stress, decreased antioxidant capacity, and enhanced cellular calcium levels are initial factors that cause endothelial cell (EC) hyperpermeability, which represents a crucial event in the pathogenesis of pre-eclampsia. Lipoxin A4 (LXA4) strongly attenuated lipopolysaccharide (LPS)-induced hyperpermeability through maintaining the normal expression of VE-cadherin and β-catenin. This effect was mainly mediated by a specific LXA4 receptor. LXA4 could also obviously inhibit LPS-induced elevation of the cellular calcium level and up-regulation of the transient receptor potential protein family C 1, an important calcium channel in ECs. At the same time, LXA4 strongly blocked LPS-triggered reactive oxidative species production, while it promoted the expression of the NF-E2 related factor 2 (Nrf2) protein. Our findings demonstrate that LXA4 could prevent the EC hyperpermeability induced by LPS in human umbilical vein endothelial cells (HUVECs), under which the possible mechanism is through Nrf2 as well as Ca2+-sensitive pathways.


INTRODUCTION
Pre-eclampsia (PE), a pregnancy-specific syndrome characterized by the onset of hypertension, proteinuria, and edema, has been reported in approximately 8% of all pregnancies [1]. Although the pathophysiology of PE still remains undefined, placental oxidative stress is regarded as a key event [1,2,3]. It is widely accepted that in pregnant women with PE, excessive generation of reactive oxidative species (ROS) in the placenta contributes to widespread maternal vascular endothelial cell (EC) hyperpermeability [2].
Lipoxins (LXs) are the first identified lipid family serving as endogenous "braking signals" in inflammation that possesses a wide spectrum of anti-inflammatory and proresolution bioactions [4]. Recently, we observed that PE patients presented much lower LX levels in systemic circulation (data not shown). In addition, we previously confirmed that LXs had an antioxidant effect both in 1057 lipopolysaccharide (LPS)-stimulated macrophages [5,6] and in carbon tetrachloride-induced liver injury [7]. Combining these above results, we speculate that deficient LX production might conduce to persistent inflammatory and oxidative conditions, which in turn cause EC hyperpermeability, but the detailed mechanism is still unknown. Therefore, the current study aimed to elucidate the role of lipoxin A 4 (LXA 4 ), a major endogenous LX, on preventing LPS-induced endothelial barrier disruption in human umbilical vein endothelial cells (HUVECs), which is a commonly applied cell model for studying systemic vascular function [8]. To further explore the possible mechanism, cellular calcium and NF-E2 related factor 2 (Nrf2), a crucial transcription factor in regulating cellular antioxidant response and phase II enzymes [9], were also investigated.

MATERIALS AND METHODS
 Materials -LXA 4 was from Cayman Chemical (USA) and was stored at -80°C until diluted in serum-free culture medium immediately before use. Commercially obtained LXA 4 was diluted in ethanol, of which the final concentration in the cell culture medium was 0.0036% or lower. Endothelial cell growth medium-2 (EGM-2) and fetal bovine serum (FBS) were purchased from GIBCO (Australia). LPS, bovine serum albumin (BSA), and Fura-2 acetoxymethyl ester (Fura-2/AM) were from Sigma Aldrich (USA  [10]. Cells at passage 3-5 were cultured in EGM and 20% FBS at 37°C with 5% CO 2 . Cells were serum starved for 24 h before experiments, and then treated with 10 µg/ml LPS with or without 100 nM LXA 4 for 12 h unless stated otherwise.  Measurement of endothelial permeability -HUVECs (1 × 10 5 /cm 2 ) were plated on 1% gelatin-coated transwell inserts in 12-well plates until confluent. Endothelial permeability was determined with permeability coefficient of albumin (Pa) according to published method [11].  Morphological observation -HUVECs were observed under an inverted microscope after regular hematoxylin and eosin (H&E) staining.  Immunofluorescence assay -Immunofluorescence assay was performed to detect VEcadherin according to our published method [12].  Western-blotting analysis -Western blotting was performed to detect VE-cadherin, βcatenin, and Nrf-2 expression as described previously [6]. with LPS with or without LXA 4 for 6 h, [Ca 2+ ]i was measured with Fura-2 according to our previous report [13]. The intensities of fluorescence due to excitation at 340 (F340) and 380 (F380) nm were measured after background subtraction, and F340/F380 was used to represent changes in [Ca 2+ ]i.  Intracellular ROS measurement -Intracellular ROS levels were determined by measuring the oxidative conversion of DCFH-DA to fluorescent compound DCFH [14]. The results were expressed as the percentage of fluorescence intensity in control cells.  Measurement of H 2 O 2 production by HUVECs -The assay is based on the detection of H 2 O 2 that reacts with Amplex Red in the presence of HRP with a 1:1 stoichiometry producing resorufin [15]. Amplex Red (50 μM) and HRP (5 U/2 ml) were added to cells exposed to LPS or LXA 4 for indicated periods of time. Fluorescence was detected at 37°C in a fluorescence spectrophotometer 30, 60, or 90 min after incubation. The excitation and emission wavelengths were 550 and 585 nm, respectively. Calibration signals were generated using known amounts of H 2 O 2 .  Statistical analysis -Statistical analysis was performed using SPSS 13.0 for Windows.
Data were presented as mean ±SEM and analyzed by Student's t test. Values of p < 0.05 indicated a statistically significant difference.

LXA 4 Prevented LPS-Induced Hyperpermeability of HUVECs
After primary HUVECs were successfully isolated and cultured, which was confirmed by CD31 and von Willebrand factor (vWF) positive staining as shown in Supplementary Fig. 1, vascular permeability was measured. As presented in Fig. 1, exposure to LPS for 12 h markedly increased the Pa value to ~183% that of control cells, while coadministration of LXA 4 attenuated this LPS-induced hyperpermeability (p < 0.05). It was also observed that 50-200 nM LXA 4 showed no dose-dependent manner. Therefore, in the following experiments, only 100 nM LXA 4 was applied.
Since the bioactions of LXA 4 were mainly elicited through the LXA 4 receptor (ALXR/FPRL-1), which had already been confirmed to exist in HUVECs [16], then 10 μM Boc-2 (an effective antagonist of FPRL-1) was used to explore the effect of FPRL-1. Results showed that when Boc-2 was administrated 30 min in advance, it effectively blocked the influence of LXA 4 on LPS-induced Pa enhancement (p < 0.01; shown in Fig.1).

LXA 4 Maintained the Normal Morphology of HUVECs
The morphology of ECs could also reflect the cell contractile state and vascular permeability, thus it was observed with routine H&E staining. As shown in Fig. 2, cells in control groups showed a cobblestonelike morphology, with tight cell-cell contact. When treated with LPS for 12 h, HUVECs presented a thinner and flatter, spindle-shaped morphology. This morphological change was concomitant with the increased albumin permeability shown in Fig. 1. Meanwhile, shapes of cells cotreated with LXA 4 were well preserved and the cell-cell conjunction was recuperated. LXA 4 alone showed no effect on the morphology of HUVECs.

LXA 4 Maintained the Expression of VE-Cadherin and β-Catenin in LPS-Stimulated HUVECs
As normal expression of VE-cadherin and β-catenin is required for the maintenance of proper endothelial adhesion junctions (AJs) and vascular permeability [17], we further determined the role of these two proteins in the protective effect of LXA 4 on EC permeability. As shown in Fig. 3, both VE-cadherin mRNA  and protein could be detectable in all of the four groups, with obviously the lowest level in cells treated with LPS (p < 0.05 vs. control cells or cells treated with both LPS and LXA 4 ). The Western blotting result was further confirmed by immunofluorescence assay with FITC-conjugated antibody to VE-cadherin. HUVECs in the control group demonstrated tight apposition between cells and strong membrane fluorescence. When LPS was administered, large intercellular gaps could be observed. At the same time, VE-cadherin staining was much lower. In cells cotreated with LXA 4 , LPS-induced changes mentioned above, like gap formation and down-regulation of VE-cadherin, were much less apparent (Fig. 4). LXA 4 alone had no significant effect on VE-cadherin expression.
It was presented in Fig. 3 that LXA 4 blocked LPS-induced β-catenin protein down-regulation (p < 0.05), while it showed no statistical effect on β-catenin protein expression in control cells.

LXA 4 Inhibited LPS-Induced [Ca2+]i Elevation and TRPC1 Expression
Calcium is critical for the maintenance of cell-cell junctions [18]; thus, we next measured [Ca 2+ ]i in HUVECs to explore the possible mechanism under which LXA 4 could maintain VE-cadherin and βcatenin expression. As shown in Fig. 5A, treatment of HUVECs with LPS for 6 h induced the increase of [Ca 2+ ]i for more than 50% that of control cells (p < 0.01). However, it was statistically lower in cells cotreated with LXA 4 (p < 0.01).
Transient receptor potential protein family C 1 (TRPC1) acts as a store-operated calcium entry channel (SOC) in the cell membrane, which regulates not only cellular calcium concentration, but also vascular permeability [19]. As shown in Fig. 5B, its mRNA expression increased obviously in cells treated with LPS, while in cells treated simultaneously by LPS and LXA 4 , this abnormal up-regulation was inhibited significantly (p < 0.05).

LXA 4 Blocked LPS-Induced ROS Generation in HUVECs
We further explored the effect of LXA 4 on cellular oxidative stress. Cells treated with LPS for 12 h showed ~1.5-fold increase of fluorescence intensity, indicating a significantly higher endogenous ROS formation (Fig. 6A). Cells cotreated with LXA 4 had significantly lower intensity compared with that in cells treated by LPS alone (p < 0.05).
The release of H 2 O 2 from HUVECs is shown in Fig. 6B. At all three time points, H 2 O 2 levels in the LPS group were much higher than those in the control group (p < 0.05) or the LPS + LXA 4 group (p < 0.05).

Effect of LXA 4 on the Antioxidant Response in HUVECs
In order to elucidate how LXA 4 inhibited ROS generation in LPS-stimulated HUVECs, Nrf2 protein level was measured. As shown in Fig. 7, it was down-regulated in LPS-treated cells, while up-regulated in cells cotreated with LXA 4 (p < 0.05).

DISCUSSION
Ever since it was found that the disappearance of Evans blue dye from plasma was faster in patients with PE [20,21], increased capillary endothelial permeability was considered as a central event in the pathogenesis of PE [22]. Despite numerous studies pointing to the potent anti-inflammatory activities of LXs, there is a paucity of research examining their direct influence on the umbilical vessel. The purpose of this study was to determine the effect of LXA 4 on modulating permeability of HUVECs and the possible mechanism. First, hyperpermeability of HUVECs was induced by administration of LPS, a commonly applied model in vascular leakage research [23]. Then it was found that LXA 4 strongly attenuated LPS-induced hyperpermeability, confirmed by both albumin permeability assay and morphological observation. ALXR is a well-known specific receptor of LXA 4 on a variety of cell types, including HUVECs [16]. In addition, Gronert et al. confirmed that LXs could bind with vascular endothelial-derived LTD4 receptor [24]. Furthermore, bioactions of LXs might also be mediated by some other distinct receptors [25]. In order to explore the signaling mechanisms responsible for the protective effect of LXA 4 , the antagonist of ALXR was applied. Results indicated that LXA 4 inhibited LPS-induced hyperpermeability in HUVECs mainly through ALXR.
Cell permeability is regulated in part by the dynamic opening and closing of cell-cell AJs. In ECs, AJs are mainly composed of VE-cadherin, an endothelium-specific member of the cadherin family, which is linked through its cytoplasmic tail to the AJ proteins p120, β-catenin, and plakoglobin [17]. In the current study, LXA 4 obviously reversed down-regulation of both VE-cadherin and β-catenin caused by LPS. This might help to explain the prevention of LXA 4 on LPS-induced hyperpermeability. Although Serhan's group previously found that LXs inhibited the neutrophil-mediated increase of vascular permeability in the mouse ear [26,27,28], and Ereso et al. recently also confirmed that LXA 4 attenuates microvascular fluid leakage during inflammation in rats [29], our study presented for the first time that LXA 4 had directly protective effect on AJs.
Regulation of endothelial permeability is a complex process that often depends on SOC [30]. Recent studies have identified TRPC1, a member of TRP family, as the essential component of the SOC in ECs [31]. The rise in [Ca 2+ ]i through TRPC1 has been established as the initial pivotal signal that precedes EC cytoskeletal reorganization and the disassembly of VE-cadherin [18,31,32]. The fact that LXA 4 obviously attenuated the expression of TRPC1 in cultured HUVECs helped us to explain how it blocked the elevation of [Ca 2+ ]i in LPS-stimulated cells. However, the details of the cellular signals on regulating TRPC1-induced Ca 2+ influx and how these signals control endothelial permeability still need to be clarified. TRPC1 is also considered to be a link between cellular Ca 2+ overload and excessive production of ROS [33]. The latter could also affect EC permeability through the regulation of VE-cadherin and βcatenin [34].
In the endeavor to explore the mechanism involved in redox disturbance of PE women, the ischemic placenta is widely regarded as a key resource of ROS [1,2,3], but it should not be ignored that activated leukocytes, neutrophils in circulation, and ECs are also likely contributors to ROS accumulation [2,3,35,36,37]. These cells could be activated by factors derived from PE placentae, such as inflammatory and vasobioactive mediators, as well as ROS [38,39]. For example, when HUVECs were stimulated with plasma from women with PE, they has increased oxidative ability [40]. Our group previously presented the inhibitory effect of LXA 4 on ROS production in several disease models [5,6,7]. Nascimento-Silva et al. further reported that an aspirin-triggered LXA 4 analog suppressed ROS generation in ECs [41]. In the current study, LXA 4 strongly blocked LPS-triggered ROS production in HUVECs.
In addition to increased pro-oxidant activity, there is also evidence for a decreased antioxidant protective capacity in women with PE [42]. Nrf2 is a crucial transcription factor in regulating cellular antioxidant response. When cells are exposed to high levels of ROS, Nrf2 translocates into the nucleus and sequentially results in up-regulation of series downstream phase II enzymes, such as NAD(P)H:quinone oxidoreductase (NQO1) and heme oxygenase 1( HO-1), which have emerged as important mediators of antioxidant and cytoprotective action [43]. Combined with the data we recently published [44], we demonstrated that LXA 4 could trigger nuclear translocation of Nrf2 and promote the expression of Nrf2, NQO1, and HO-1. It has illuminated a possible mechanism of its antioxidant effect on HUVECs.
In summary, the present study indicates for the first time that LXA 4 , at physiological concentrations, could prevent the LPS-induced EC hyperpermeability through maintaining the expression of VE-cadherin and β-catenin. It might involve modulation of LXA 4 on Ca 2+ and redox homeostasis. Although there are still deep gaps to fill, especially about the underlying mechanism, our data might also contribute to the potential therapeutic value of LXs for PE and other vascular endothelial hyperpermeability-associated diseases.