Internalization and Transportation of Endothelial Cell Surface KCa2.3 and KCa3.1 in Normal Pregnancy and Preeclampsia

Altered redox state modulates the expression levels of endothelial KCa2.3 and KCa3.1 (KCas) in normal pregnancy (NP) and preeclampsia (PE), thereby regulating vascular contractility. The mechanisms underlying KCas endocytosis and transportation remain unknown. We investigated the regulation of KCas expression in plasma membrane (PM) during NP and PE. Cultured human uterine artery endothelial cells were incubated in serum from normal nonpregnant women and women with NP or PE, or in oxidized LDL-, or lysophosphatidylcholine- (LPC-) containing a medium for 24 hours. NP serum elevated PM levels of KCas and reduced caveolin-1 and clathrin levels. PE serum, oxidized LDL, or LPC reduced PM levels of KCas and elevated caveolin-1, clathrin, Rab5c, and early endosome antigen-1 (EEA1) levels. Reduced KCas levels by PE serum or LPC were reversed by inhibition of caveolin-1, clathrin, or EEA1. Catalase and glutathione peroxidase 1 (GPX1) knockdown elevated PM-localized KCas levels and reduced caveolin-1 and clathrin levels. Elevated KCa2.3 levels upon catalase and GPX1 knockdown were reversed by PEG-catalase treatment. An H2O2 donor reduced clathrin and Rab5c. In contrast, elevated clathrin, caveolin-1, or colocalization of caveolin-1 with KCa3.1 by PE serum or LPC was reversed by NADPH oxidase inhibitors or antioxidants. A superoxide donor xanthine+xanthine oxidase elevated caveolin-1 or Rab5c levels. We concluded that KCas are endocytosed in a caveola- or a clathrin-dependent manner and transported in a Rab5c- and EEA1-dependent manner during pregnancy. The endocytosis and transportation processes may slow down via H2O2-mediated pathways in NP and may be accelerated via superoxide-mediated pathways in PE.


Introduction
K Ca 2.3 and K Ca 3.1 play an important role in endothelial control of vascular contractility. Activation of these K + channels induces K + efflux and endothelial hyperpolarization, which hyperpolarize vascular smooth muscle cells (VSMCs) by activating inward-rectifier K + channels and spreading to VSMCs through gap junctions, respectively [1][2][3]. In addition, endothelial hyperpolarization enhances Ca 2+ entry through Ca 2+permeable channels such as transient receptor potential channels by increasing its electrical driving force and elevates intracellular Ca 2+ levels [4], which stimulates nitric oxide (NO) production in endothelial cells (ECs) [5]. NO and VSM hyperpolarization relax blood vessels, thereby controlling vascular contractility. The contribution of NO and VSM hyperpolarization to the control of vascular contractility might vary between conduit arteries and resistant arteries. The contribution of NO was most prominent in the aorta, whereas that of VSM hyperpolarization was most prominent in the distal mesenteric arteries, suggesting that VSM hyperpolarization plays a more important role in the control of vasorelaxation in resistant arteries than in conduit arteries [6][7][8][9]. Since resistant arteries are the main regulators of systemic vascular resistance, endothelial K Ca 2.3 and K Ca 3.1 might play an important role in the regulation of blood pressure.

Human Subjects.
The study population consisted of Asian women who were not pregnant or had either a NP or PE (Table 1). Pregnancies were considered normal when patients did not have medical and obstetric complications of pregnancy and delivered a newborn at a gestational age of 37-42 weeks. Preeclampsia was defined by a systolic blood pressure over 140 mmHg and a diastolic blood pressure over 90 mmHg after 20 weeks of gestation in a previously normotensive woman, and the onset of proteinuria exceeding 300 mg of protein during 24 hours of urine collection. Nonpregnant women were healthy premenopausal volunteers taking no medications. Preeclamptic patients and normal pregnant women were matched for age (±3 years) and gestational age (±2 weeks), and nonpregnant healthy female volunteers were matched for age (±3 years). Blood samples were obtained from subjects during the third trimester of pregnancy. The study population was monitored at the Department of Obstetrics and Gynecology from the first trimester until their pregnancy was completed without complications. Exclusion criteria included the following: altered renal function, diabetes or chronic diseases, twin pregnancies, recurrent miscarriages, fetal growth retardation, and abruptio placenta.
Smokers and women with a history of essential hypertension were also excluded from this study. Gestational age was defined as the interval between the first day of the mother's last menstrual period and the date of delivery.
2.3. Cell Culture and Serum Treatment. Human uterine microvascular ECs (HUtMECs), which were purchased from PromoCell GmbH (Heidelberg, Germany), were maintained in EC Growth Medium MV2 (PromoCell GmbH). For serum treatment, HUtMECs were plated in 6-well plates for 24 hours. The concentration of fetal bovine serum in a culture medium was gradually decreased from 10% to 5, 2, and 0% over 30 minutes, and HUtMECs were incubated in a serum-free medium for 30 minutes. After that, a culture Values shown are mean ± SEM and exclusively composed of plasma donors. medium was substituted with serum from normal nonpregnancy (NNP) women or women with NP or PE, and the cells were incubated for 24 hours. Mouse aortic endothelial cells (MAECs) were isolated from the mouse aortas as described [16]. Briefly, periadventitial fats and connective tissues around the aorta were carefully cleaned in Ca 2+ -free phosphate-buffered saline under a dissecting microscope. Matrigel (BD Biosciences, San Jose, CA) was plated and polymerized at 37°C for 30 minutes. After that, aorta pieces were placed with the intima side down on the Matrigel. To demonstrate the endothelial nature of the cell, 1,1 ′ -dioctadecyl-3,3,3 ′ ,3 ′ -tetramethyl-indocarbocyanine perchlorate-labeled acetylated low-density lipoprotein (Biomedical Technologies Inc., Stoughton, MA) uptake assay was employed. MAECs were used within 2 passages and not above 3 passages.
2.4. Immunoblotting and Immunoprecipitation. For immunoblotting, cell lysates were used to examine the protein level. After proper processing of each type of sample, total protein was measured using the bicinchoninic acid assay (Pierce Biotechnology, Rockford, IL). The same amount of total protein was analyzed using SDS-PAGE on 7.5-12% gels and transferred to nitrocellulose membrane (Invitrogen, Eugene, OR). Membranes were blocked for 1 hour in 5% bovine serum albumin in Tris-buffered saline with 0.1% Tween-20 and incubated overnight at 4°C with primary antibodies (Abs) diluted in blocking buffer. Membranes were then washed three times with Tris-buffered saline with 0.1% Tween-20 and incubated for 1 hour with horseradish peroxidase-conjugated secondary Abs diluted in blocking buffer. The immunoblots were visualized by chemiluminescence reagents bought from GE Healthcare (Piscataway, NJ). Data processing was performed using a luminescent image analyzer LAS-3000 (Fujifilm, Tokyo, Japan) and IMAGE GAUSE software.
For immunoprecipitation, cells were washed twice with phosphate buffer saline and lysed in lysis buffer containing protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO) at 4°C for 1 hour. The lysate samples were prepared by centrifugation at 12000 × g for 30 minutes to eliminate the cell debris. Total protein concentration was estimated as described above. The lysate samples were precleared with a nonspecific IgG Ab. Total 30 μL of 50% protein G-coupled dynabead slurry (Invitrogen) was added to an Ab (1-5 μg) diluted in 200 μL phosphate buffer saline with Tween 20, incubated for 15 minutes at room temperature (about 20°C) with rotating, and washed. Precleared lysate samples were incubated with the dynabeads-Ab by rotating at 4°C overnight or at room temperature for 30 minutes. Following that, the dynabeads-antigen-Ab complexes were washed three times with the lysis buffer, and the antigen was eluted in 2x SDS-PAGE sample buffer by heating at 70°C for 10 minutes. Immunoprecipitates were separated using 7.5-12% SDS-PAGE and analyzed by immunoblotting.
2.5. Biotinylation of Cell Surface Protein. Cells were washed twice with phosphate buffer saline and labeled with 1 mM sulfosuccinimidyl-2-(biotinamido)ethyl-1, 3-dithiopropionate (EZ-Link-sulfo-NHS-SS-biotin; Pierce Biotechnology) in labeling buffer (150 mM NaCl, 20 mM HEPES, 3 mM CaCl 2 , and 1 mM MgCl 2 ) for 30 minutes to 1 hour at room temperature. After the cells were washed, any nonreacted biotinylation reagent was quenched with 100 mM glycine, the cells were lysed in NP40 lysis buffer, and the proteins were incubated with 30 μL of dynabead M-280 streptavidin (Invitrogen) for 3 hours at 4°C with rotation and then washed three times with lysis buffer. The proteins were eluted from streptavidin bead in 2x SDS-PAGE sample buffer by heating at 65°C for 5 minutes. Supernatant was subjected to 7.5-12% SDS-PAGE and analyzed by immunoblotting.
2.6. siRNA Transfection. Negative control siRNAs (SN-1012) and siRNAs against EEA1 (SDH-1001) were purchased from Bioneer (Daejeon, Korea). Negative control siRNAs (sc-36869) and siRNAs against NADPH oxidase 4 (NOX4; sc-41586) were purchased from Santa Cruz Biotechnology. ECs were transiently transfected with the siRNAs using an siRNA transfection reagent (Santa Cruz Biotechnology) according to the procedure suggested by the manufacturer. Cell lysates were prepared 24 hours after transfection, and immunoblotting was performed using anti-EEA1 Abs.

2.7.
Electrophysiology. The patch-clamp technique was used in whole-cell configurations at 20-22°C. Whole-cell currents were measured using ruptured patches and monitored in voltage-clamp modes with an EPC-9 (HEKA Elektronik, Lambrecht, Germany). The holding potential was 0 mV, and currents were monitored by the repetitive application of voltage ramps from −100 to +100 mV with a 10-second interval (sampling interval 0.5 milliseconds, 650 millisecond duration). The standard external solution contained (in mM) 150 NaCl, 6 KCl, 1.5 CaCl 2 , 1 MgCl 2 , 10 HEPES, and 10 glucose, pH adjusted to 7.4 with NaOH. The pipette solution for whole-cell recording contained (in mM) 40 KCl, 100 K-aspartate, 2 MgCl 2 , 0.1 EGTA, 4 Na 2 ATP, and 10 HEPES, pH adjusted to 7.2 with KOH. To buffer free Ca 2+ , the appropriate amount of Ca 2+ (calculated using CaBuf software; G. Droogmans, Leuven, Belgium) was added in the presence of 5 mM EGTA. K Ca 3.1 currents were activated by loading 1 μM Ca 2+ via a patch pipette in whole-cell clamped MAECs. K Ca 3.1 current was normalized to cell capacitance, and the selective K Ca 3.1 blocker TRAM-34-sensitive current was measured as the K Ca 3.1 current.
The final concentration of DMSO, chloroform, or methanol in media was less than 0.1%, and these solvents did not have any effect on the experiments tested in this study (data not shown).
2.9. Statistics. Data represent mean ± SEM. To prove the statistical significance between groups, one-way ANOVA with Bonferroni's post hoc or 2-tailed Student's t-test was used. A P value of 0.05 or lower was considered statistically significant. Calculations were performed with SPSS 14.0 for Windows (SPSS, Chicago, IL).

Endothelial Membrane Levels of K Ca 2.3 and K Ca 3.1 Are
Altered during Pregnancy. We compared the effects of soluble serum factors from NNP and NP women on PM-localized K Ca 2.3 or K Ca 3.1 levels by incubating HUtMECs in serum from NNP and NP women for 24 hours. K Ca 2.3 or K Ca 3.1 proteins were biotinylated at the cell surface and labeled with horseradish peroxidase-conjugated streptavidin. The levels of PM-localized K Ca 2.3 or K Ca 3.1 were significantly higher in ECs treated with NP serum than in ECs treated with NNP serum (Figure 1(a)). We then compared the effects of serum from NP and PE women on PM levels of K Ca 2.3 or K Ca 3.1 in HUtMECs. The levels of PM-localized K Ca 2.3 or K Ca 3.1 proteins were significantly lower in ECs treated with PE serum than in ECs treated with NP serum (Figure 2(b)). In addition, we compared the effect of serum from NNP, NP, and PE women on PM levels of K Ca 2.3 ( Figure 1(c)) or K Ca 3.1 (Figure 1(d)) in HUtMECs. The levels of PM-localized K Ca 2.3 or K Ca 3.1 were not changed in ECs treated with NNP serum, compared to those treated with normal culture medium (CM), and were elevated in ECs treated with NP serum, compared to those treated with NNP serum or CM. Increases in the K Ca 2.3 or K Ca 3.1 levels were significantly reduced in ECs treated with PE serum than in ECs treated with NP serum. Since oxidized LDL is among the causative factors to induce endothelial dysfunction in PE, we examined the effects of oxidized LDL on PM levels of K Ca 2.3 and K Ca 3.1 by incubating HUtMECs in oxidized LDL containing a culture medium for 24 hours. We found that the levels of PM-localized K Ca 2.3 or K Ca 3.1 were significantly reduced upon incubation (Figure 1(e)). Reduced K Ca 2.3 or K Ca 3.1 levels by oxidized LDL treatment were reversed by blocking the oxidized LDL receptors using an anti-LOX1 Ab (Figure 1(e)). These results suggest that the expression and localization of K Ca 2.3 or K Ca 3.1 in PM are elevated in NP compared to that in NNP, and the elevation was attenuated in PE.

Caveolae and Clathrin
Are Involved in the Internalization of K Ca 2.3 and K Ca 3.1. The levels of PM-localized proteins, such as ion channels, can be modulated by caveoladependent internalization. Caveolins, the essential structural elements of caveolae, are suggested to be scaffolding proteins that facilitate the compartmentalization of various signaling molecules or proteins within caveolae. We thus investigated whether caveola-dependent internalization of K Ca 2.3 or K Ca 3.1 occurs during pregnancy by examining Cav-1 levels. Cav-1 levels were markedly lower in the ECs treated with NP serum than in ECs treated with NNP serum, and VEGF receptor (VEGFR) inhibition using an anti-VEGFR1 Ab or anti-VEGFR2 Ab enhanced Cav-1 levels in ECs treated with NP serum (Figure 2(a)), indicating that NP serum decreases Cav-1 levels via VEGFR activation. In contrast, Cav-1 levels were markedly higher in ECs treated with PE serum than in ECs treated with NP serum (Figure 2(b)). In addition, LPC, the major component of oxidized LDL, enhanced Cav-1 levels in a concentration-dependent manner (Figure 2(b)). To confirm the presence of Cav-1 in the inner leaflet of the PM, the cell surface was biotinylated with a membrane impermeable agent (NHS-SS-biotin). Biotinylated PM proteins from whole cell lysates were isolated on a streptavidin column and, following elution and SDS-PAGE, were blotted for Cav-1 with anti-Cav-1 Ab. PM localization of Cav-1 was significantly elevated upon treatment with oxidized LDL (Figure 2(c)). We then examined whether caveolae are involved in the regulation of the levels of K Ca 2.3 or K Ca 3.1. Colocalization of K Ca 3.1 with Cav-1 was examined using coimmunoprecipitation. Colocalization of K Ca 3.1 with Cav-1 was markedly higher in ECs treated with PE serum than in ECs treated with NP serum (Figure 2(d)). Cav-1 inhibition using the Cav-1 inhibitor MβCD elevated K Ca 2.3 ( Figure 2(e)) or K Ca 3.1 (Figure 2(f)) levels in ECs treated with PE serum or LPC. These results suggested that caveoladependent internalization is involved in regulating the PM localization of K Ca 2.3 or K Ca 3.1 during pregnancy. Caveoladependent internalization process might be delayed in NP, whereas it might be facilitated in PE.
Previously, we reported that K Ca 3.1 proteins on the PM are internalized via a clathrin-dependent process in Fabry disease [12]. We, therefore, examined whether clathrindependent internalization is involved in regulating PM localization of K Ca 2.3 or K Ca 3.1 during pregnancy. Clathrin levels were not altered in ECs treated with NNP or NP serum (Figure 3(a)). In contrast, clathrin levels were markedly higher in the ECs treated with PE serum than in ECs treated with the NP serum (Figure 3(b)). Then, clathrin protein was biotinylated at the cell surface. PM clathrin levels were significantly higher in ECs treated with PE serum than in ECs treated with NP serum (Figure 3(c)). Colocalization of K Ca 3.1 with clathrin was markedly enhanced in ECs treated with PE serum than in ECs treated with NP serum, and LPC markedly increased colocalization of K Ca 3.1 with clathrin (Figure 3(d)). Furthermore, the clathrin inhibitor chlorpromazine enhanced K Ca 2.3 (Figure 3(e)) and K Ca 3.1 levels (Figure 3(f)) in ECs treated with PE serum or LPC. These results suggest that K Ca 2.3 and K Ca 3.1 are internalized via a clathrin-dependent process. Clathrin-dependent internalization process might not be affected in NP, whereas it might be facilitated in PE.

A Rab5c
-and EEA1-Dependent Process Mediates K Ca 2.3 or K Ca 3.1 Degradation. The small GTPases Rab5 and EEA1 play a rate-limiting role in membrane docking or fusion in the early endocytic pathway [17,18], and our previous study suggested that K Ca 3.1 is transported into early endosomes via a Rab5c-and EEA1-dependent process in Fabry disease [12]. We thus examined whether a Rab5c-and EEA1-dependent process mediates K Ca 2.3 and K Ca 3.1 transportation during pregnancy. Compared to ECs treated with NNP serum, levels of EEA1 (Figure 4(a)) or Rab5c (Figure 4(b)) were slightly decreased in the ECs treated with NP serum. However, significant difference was not found in ECs treated with NNP     LPC increased Rab5c and EEA1 levels in a concentrationdependent manner (Figure 4(d)). We examined whether Rab5c is involved in the regulation of the levels of K Ca 3.1 using coimmunoprecipitation. LPC enhanced colocalization of Rab5c with K Ca 3.1 (Figure 4(e)). We then examined the effect of EEA1 inhibition on LPC-induced downregulation of K Ca 2.3 or K Ca 3.1 by using siRNA against EEA1. K Ca 2.3 (Figure 4(f)) and K Ca 3.1 (Figure 4(g)) levels, which were reduced by PE serum or LPC, recovered upon EEA1 inhibition. These results suggest that Rab5c and EEA1 are involved in K Ca 2.3 or K Ca 3.1 downregulation in PE.  [10]. In addition, K Ca 3.1 levels were elevated in mouse aortic endothelial cells (MAECs) from catalase/GPX1 double knockout (catalase -/-/GPX1 -/-) mice [11]. Thus, we examined whether catalase and GPX1 knockdown affects PM levels of K Ca 2.3 or K Ca 3.1 using wild-type and catalase -/-/GPX1 -/mice. PM-localized K Ca 2.3 or K Ca 3.1 levels were markedly increased in catalase -/-/GPX1 -/-MAECs, compared to wildtype MAECs (Figure 5(a)). We then compared Cav-1 or clathrin levels in wild-type and catalase -/-/GPX1 -/-MAECs. In catalase -/-/GPX1 -/-MAECs, levels of Cav-1 ( Figure 5(b)) or clathrin ( Figure 5(c)) were markedly reduced, and K Ca 2.3 levels were elevated ( Figure 5(c)). Thus, inverse relation between K Ca 2.3 and clathrin levels was found. The increase in K Ca 2.3 levels seen in these cells was reversed by treatment with polyethylene glycol-(PEG-) catalase ( Figure 5(d)). Furthermore, a H 2 O 2 donor TBHP reduced clathrin ( Figure 5(e)) and Rab5c ( Figure 5(f)) levels in a concentration-dependent manner. These results suggest that catalase and GPX1 downregulation slows down the internalization and trafficking of K Ca 2.3 and K Ca 3.1 from PM via H 2 O 2 -mediated pathways.
On the other hand, we had previously reported that NADPH oxidase 2 (NOX2) and NOX4 upregulation and SOD downregulation enhance superoxide levels, thereby downregulating K Ca 2.3 and K Ca 3.1 in PE [10,19]. We thus examined whether superoxide facilitates the internalization and transportation processes of K Ca 2.3 and K Ca 3.1 in PE. Compared to cells treated with NP serum, PM-localized clathrin levels were markedly elevated in ECs treated with PE serum, and the elevation was reversed by the treatment with NOX4 siRNA (Figure 6(a)). LPC elevated Cav-1 levels, and the elevation was reversed by the treatment with the antioxidants, tempol or tiron (Figure 6(b)). LPC increased colocalization of K Ca 3.1 with Cav-1, clathrin, or Rab5c, which was inhibited by a pan-NOX inhibitor VAS2870 or apocynin (Figure 6(c)). Furthermore, a superoxide donor X/XO increased Cav-1 ( Figure 6(d)) or Rab5c (Figure 6(e)) levels in ECs in a concentration-dependent manner. These results suggest that the internalization and transportation processes of K Ca 2.3 and K Ca 3.1 facilitated by superoxide contribute to the downregulation of K Ca 2.3 and K Ca 3.1 in PE.

Discussion
In this study, we observed for the first time that PM-localized K Ca 2.3 and K Ca 3.1 are internalized via caveola-or clathrindependent processes and transported via a Rab5c-and EEA1-dependent process in ECs (Figure 7). Compared to NNP, the internalization and transport processes are delayed in NP, thus elevating PM-localized K Ca 2.3 and K Ca 3.1 levels. However, compared to NP, the internalization and transport processes are facilitated in PE, thus reducing PM-localized K Ca 2.3 and K Ca 3.1 levels. Soluble factors in PE serum, such as oxidized LDL, might induce internalization of K + channel proteins from the PM via clathrin-or caveola-dependent processes and thereby attenuate the pregnancy-associated K Ca 2.3 and K Ca 3.1 upregulation. Since endothelial K Ca 2.3 and K Ca 3.1 play important roles in the control of vascular contractility, altered levels of these K + channels might explain the hemodynamic changes seen during the progress of pregnancy, both NP and PE. K Ca 2.3 was internalized from PM via a caveola-or clathrin-dependent process. The involvement of caveolae in the internalization of K Ca 2.3 is consistent with the finding that K Ca 2.3 is observed in caveolae, Cav-1-rich membrane fractions [20], and that the endocytosis of K Ca 2.3 from the cell membrane is dependent upon both Cav-1 and dynamin II [21]. In addition, we showed that clathrin is involved in K Ca 2.3 internalization. Internalized K Ca 2.3 might be transported via a EEA1-and Rab5-dependent process, since LPC-induced reduced K Ca 2.3 levels were reversed by siRNA against EEA1 (Figure 4(f)). Similarly, Gao et al. demonstrated the involvement of Rab5-containing endosome in the endocytosis of K Ca 2.3 from PM [21]. These mecha-nisms involved in K Ca 2.3 internalization and transportation were similar to those for K Ca 3.1 internalization and transportation.
Previously, we demonstrated that PM K Ca 3.1 proteins are internalized via a clathrin-dependent process and transported in a Rab5c-and EEA1-dependent process in Fabry disease [12]. In addition, we demonstrated that caveolae are involved in K Ca 3.1 internalization from cell membrane, although endothelial K Ca 3.1 is suggested to be present in noncaveolar membrane fractions [20,22]. However, K Ca 3.1, which is not associated with Cav-1 under baseline conditions, was found to colocalize with Cav-1 during shear stress conditions [23]. In addition, PE serum induced colocalization of K Ca 3.1 with Cav-1 (Figure 2(d)). Thus, stimulation by serum components or shear stress might lead to colocalization of K Ca 3.1 and Cav-1. In addition, clathrin-dependent internalization followed by Rab5c-and EEA1-dependent transportation was found to regulate membrane K Ca 3.1 levels during pregnancy.
Altered redox state might be involved in the regulation of internalization and transportation of K Ca 2.3 and K Ca 3.1 in NP and PE. We previously showed that H 2 O 2 levels are elevated via downregulation of catalase and GPX1 during NP