Increased Soluble Epoxide Hydrolase in Human Gestational Tissues from Pregnancies Complicated by Acute Chorioamnionitis

Chorioamnionitis (CAM) is primarily a polymicrobial bacterial infection involving chorionic and amniotic membranes that is associated with increased risk of preterm delivery. Epoxyeicosatrienoic acids (EETs) are eicosanoids generated from arachidonic acid by cytochrome P450 enzymes and further metabolized mainly by soluble epoxide hydrolase (sEH) to produce dihydroxyeicosatrienoic acids (DHETs). As a consequence of this metabolism of EETs, sEH reportedly exacerbates several disease states; however, its role in CAM remains unclear. The objectives of this study were to (1) determine the localization of sEH and compare the changes it undergoes in the gestational tissues (placentas and fetal membranes) of women with normal-term pregnancies and those with pregnancies complicated by acute CAM; (2) study the effects of lipopolysaccharide (LPS) on the expression of sEH in the human gestational tissues; and (3) investigate the effect of 12-(3-adamantan-1-yl-ureido)-dodecanoic acid (AUDA), a specific sEH inhibitor, on LPS-induced changes in 14,15-DHET and cytokines such as interleukin- (IL-) 1β and IL-6 in human gestational tissues in vitro and in pregnant mice. We found that women with pregnancies complicated by acute CAM had higher levels of sEH mRNA and protein in fetal membranes and villous tissues compared to those in women with normal-term pregnancies without CAM. Furthermore, fetal membrane and villous explants treated with LPS had higher tissue levels of sEH mRNA and protein and 14,15-DHET than those present in the vehicle controls, while the administration of AUDA in the media attenuated the LPS-induced production of 14,15-DHET in tissue homogenates and IL-1β and IL-6 in the media of explant cultures. Administration of AUDA also reduced the LPS-induced changes of 14,15-DHET, IL-1β, and IL-6 in the placentas of pregnant mice. Together, these results suggest that sEH participates in the inflammatory changes in human gestational tissues in pregnancies complicated by acute CAM.


Introduction
Chorioamnionitis (CAM) is primarily a polymicrobial bacterial infection involving chorionic and amniotic membranes and amniotic fluid and is implicated in 0.5% to 10% of all pregnancies [1]. A number of studies have demonstrated that CAM increases the risk of preterm labor and subsequent preterm delivery [2,3]. In addition, CAM is associated with adverse neonatal outcomes including cerebral palsy and bronchopulmonary dysplasia [4,5].
Eicosanoids are a group of lipid mediators generated from arachidonic acid (AA) by the activity of cyclooxygenases (COX), lipoxygenases (LOX), and cytochrome P450 (CYP450) enzymes. Previous studies-mostly focusing on the COX and LOX pathways-have identified the essential roles of AA metabolism and eicosanoids in the pathophysiology of CAM. These include increased levels of AA in the placentas of women with CAM and preterm delivery compared to women with preterm delivery but without CAM [6]; upregulation of COX-2 and prostaglandin (PG) G/H synthase, along with certain inflammatory genes, in the amnion and choriodecidua of women with CAM [7]; increased production of PGE 2 , PGF 2α , and leukotriene B 4 (LTB 4 ) from the placentas and fetal membranes in women with CAM than in those without CAM [8,9]; and greater concentrations of PGs and LTB 4 in the amniotic fluid of women with intra-amniotic infections than in women without intra-amniotic infections [10][11][12]. Nevertheless, the role of AA metabolism via the CYP450 pathway in human gestational tissues of pregnancies complicated by acute CAM remains unclear.
Unlike most metabolites of the COX and LOX pathways, metabolites of the CYP450 enzymes, mainly epoxyeicosatrienoic acids (EETs) and dihydroxyacids, have primarily vasodilatory, antihypertensive, anti-inflammatory, and natriuretic properties [13,14]. EETs are further metabolized by soluble epoxide hydrolase (sEH), which adds water across the epoxide to give the corresponding dihydroxyeicosatrienoic acids (DHETs) [14,15]. As a consequence of its metabolism of EETs, sEH reportedly contributes significantly to the pathogenesis of several disease states such as diabetes, hypertension, and pain [16]. Indeed, increasing evidence indicates that the inhibition of sEH increases levels of EETs, which have anti-inflammatory effects and can prevent the development of hypertension, atherosclerosis, heart failure, fatty liver, and multiple organ fibrosis [14,16,17].
EETs, DHETs, and sEH have been detected in many human tissues including the placentas and fetal membranes [18,19]. Considering its central role in the metabolism of EETs, we surmised that there is dysregulation of sEH in human gestational tissues (placental villi and fetal membranes) in pregnancies complicated by acute CAM. Therefore, the objectives of this study were to (1) determine and compare the localization and changes of sEH in human gestational tissues between women with normal-term pregnancies and those with pregnancies complicated by acute CAM; (2) study the effects of lipopolysaccharide (LPS)-an endotoxin found in gram-negative bacteria that is reportedly elevated in the amniotic fluid and choriodecidua during acute CAM-on the expression of sEH in human gestational tissues in vitro; and (3) investigate the effect of 12-(3-adamantan-1-yl-ureido)-dodecanoic acid (AUDA)-a specific sEH inhibitor-on LPS-induced changes in 14,15-DHET and cytokines such as interleukin-1β and IL-6 in human gestational tissues in vitro and in pregnant mice.

Materials and Methods
Conduction of this study was approved by the Institutional Review Board of Chang Gung Memorial Hospital (No. 201601866B0 and No. 201802304B0). All placental samples were collected after the subjects enrolled herein provided written informed consent for the use of the samples. Unless otherwise indicated, the reagents used in the study were purchased from Sigma-Aldrich (St. Louis, MO, USA).

Placental Collection and Sampling.
Human placentas and attached fetal membranes were collected from 20 normalterm pregnant women experiencing cesarean delivery due to previous section or fetal malpresentation and 16 women with cesarean delivery for acute CAM. We used the following criteria to identify CAM in the subjects of this study: maternal fever (body temperature higher than 38°C), the rupture of membranes, and the presence of one of the following conditions: leukocytosis (white blood cell counts higher than 12,000/μL), elevated serum levels of C-reactive protein (>5 mg/dL), and fetal tachycardia (a baseline fetal heart rate > 160 beats/min) followed by histological confirmation (defined as extensive infiltration of neutrophils from the subchorionic space throughout the chorion) [20].
After the placenta was delivered, fetal membranes (amnion and choriodecidua) were separated from the placenta by blunt dissection under sterile conditions. Furthermore, villous tissues were randomly sampled from five distinct sites from the maternal side using a transparent sheet bearing a systematic array of sampling windows; each site was middistance between the cord insertion and the periphery of the placenta and midway between the chorionic and basal plates. After a brief rinse in cold phosphate-buffered saline, several tissue fragments of fetal membrane rolls and villous samples were either fixed in 4% paraformaldehyde or snap frozen in liquid nitrogen. The remaining fragments were placed in culture medium (medium-199 with 25 mM HEPES, Earle's salts, and L-glutamine supplemented with 5% heat-inactivated fetal bovine serum, 100 U/mL penicillin, 100 μg/mL streptomycin, and 0.25 μg/mL amphotericin B) and transferred to the laboratory on ice for individual experiments. All placental samples were collected and processed within 10 minutes of the delivery.

Fetal Membrane and Villous Explant
Culture. Fetal membrane and villous explant cultures were performed as described previously [21,22]. Briefly, under sterile laboratory conditions, membranes were checked for integrity to ensure that the amnion and chorion remained attached to one another. The membranes were rinsed with culture medium to remove residual adherent blood, cut into approximately 6 × 6 cm 2 pieces, bound together with Orthodontic Elastic Rubber Bands (5/16 inch; ORMCO Z-pak elastics, Ormco Corporation, Glendora, CA, USA), and placed on the upper chamber of a Corning Transwell® device (24 mm diameter without synthetic membrane; Corning, NY, USA). The inserts were placed in 6-well culture plates to create two distinct chambers containing 2 mL of culture medium on both sides of the mounted membranes. The amniotic side of the membranes faced the inner chamber of the Transwell® device; excess tissue extending beyond the elastic band was trimmed with a scalpel.
For the preparation of villous explants, villous samples were dissected into pieces weighing 5-10 mg and placed in culture medium; six to eight such pieces were suspended at the gas-liquid interface in individual Costar Netwell® inserts (24 mm diameter, 500 μm mesh) suspended in 3 mL of culture medium. Both fetal membranes and villous explants were equilibrated in culture media at 37°C and 5% CO 2 in a humidified incubator overnight. The medium was refreshed the following day, and the explants were cultured under individual experimental conditions.

Explant
Culture with or without LPS and AUDA Administration. Based on our preliminary concentrationresponse experiments, LPS from Escherichia coli 026:B6 (Sigma-Aldrich) at a concentration of 10 μg/mL and AUDA (Cayman Chemical, Ann Arbor, MI, USA) at a concentration of 10 μM were used to assess their effects on the changes in sEH and 14,15-DHET in tissue samples and the production of cytokines including interleukin-(IL-) 1β and IL-6 in the medium for explant culture experiments ( Supplementary  Figures 1 and 2). After overnight rest, fetal membrane and villous explants were incubated in fresh media with or without LPS in the absence or presence of AUDA for 24 hours. Thereafter, the medium was recovered from both sides of the membranes from the two-chamber culture system or from the villous explant cultures and stored at -80°C. After collection of medium, amnion, choriodecidua, and villous samples were removed from the Transwell® frame and snap frozen in liquid nitrogen for subsequent analysis. Ten placentas were used to study the effect of LPS on the transcription and translation of sEH in fetal membranes and villous explants, and eight were used to study the effect of AUDA on LPS-induced changes in the expression of 14,15-DHET in the tissue homogenates and IL-1β and IL-6 in the media of explant cultures, respectively.

Animal Experiments.
Time-pregnant C57BL/6 mice ordered from BioLASCO (Taipei, Taiwan) were used in this study. Mice were allowed free access to water and maintained on a 12 h/12 h dark cycle at a controlled temperature (22-25°C) and humidity (40-60%). All the experimental protocols were approved by the Institutional Animal Care and Use Committee at Cheng Hsin General Hospital (animal permit number CHGH-105-10), and all animals were cared for in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996).
On day 15 gestation, pregnant mice received a single intraperitoneal injection of LPS (300 μg/kg in 0.1 mL saline) derived from Escherichia coli O111:B4. An equivalent saline injection was administered for the control group (sham control). Ten minutes after LPS administration, mice were randomized to receive an intraperitoneal injection of either AUDA at a concentration of 20 mg/kg or an equal volume of vehicle (5% DMSO). Six hours after LPS administration, pregnant mice were terminally anaesthetized using 5% isoflurane (2 L/min O 2 flow rate), and the uterus was exteriorized via a midline laparotomy to identify individual gestational sacs. The gestational sacs were then opened, and individual placentas were collected, snapped frozen in liquid nitrogen, and stored at -80°C for further assessment. Eight mice for each group were used for this experiment.
2.5. Tissue Integrity Assessment. Fresh villous tissues and membranes from five different placentas were incubated in phenol red-free culture medium to monitor tissue integrity after LPS stimulation with or without AUDA, as assessed by the release of lactate dehydrogenase into the culture medium, using a commercial kit (Cytotoxicity Detection Kit, Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer's instructions.
2.6. Immunohistochemistry. Immunohistochemistry for sEH, cytokeratin 7 (CK7), and CD68 was performed as previously described [20]. Briefly, after quenching endogenous peroxidase activity and blocking nonspecific binding, sections were reacted with mouse anti-human sEH (A-5) monoclonal antibody (1 : 10 dilution; catalog no. SC-166961; Santa Cruz Biotechnology, Inc. Dallas, TX, USA), anti-human CK7 monoclonal antibody (1 : 1000 dilution, clone OV-TL 12/30, Dako Omnis, Agilent, Santa Clara, CA, USA), or antihuman CD68 monoclonal antibody (1 : 1000 dilution, clone KP1, Dako Omnis) at 4°C overnight. Further processing for colorimetric detection was according to the instructions for the VECTASTAIN Elite ABC Kit (Vector Laboratories, Burlingame, CA, USA) using diaminobenzidine as the peroxidase substrate. The specificity of a staining reaction was assessed in several control procedures, including omission of the primary antibody or substitution of the primary antibody with nonimmune mouse isotypic IgG. Tissue sections from C57BL/6 mice livers were used as the positive controls for sEH immunostaining [23]. Sections were viewed and photographed under a differential interference contrast microscope (Nikon Eclipse 80i, Nikon Corporation, Tokyo, Japan).

Western Blot.
Western blotting was performed as previously detailed [24]. Fifty to 100 micrograms of cytosolic proteins was separated by 8-12% SDS-PAGE, transferred to  Figure 1: Immunoreactivity for sEH in fetal membranes from normal pregnant women and women with pregnancies complicated by acute CAM. Immunostaining of CK7 delineates the localization of amniotic epithelium and trophoblast cells in choriodecidua (a, b). Compared to normal pregnancy, fetal membranes from women with acute CAM had increased infiltration of monocytes/macrophages, as demonstrated by positive anti-CD68 immunostaining, within the amnion and choriodecidua (d, e). In normal pregnancy, immunostaining of sEH was observed in the amniotic epithelium (g). The immunostaining of sEH was generally more intense, both in the amnion and choriodecidua, in women with acute CAM compared with that in normal pregnant women (h). There was no staining observed in the negative controls when the primary antibody was substituted with nonimmune IgG (c, f, and i). Scale bar = 50 μm. AE, amniotic epithelium; CH, choriodecidua; arrows, endothelium.
nitrocellulose membranes, and probed with the mouse antihuman sEH (A-5) monoclonal antibody (1 : 200 dilution; catalog no. SC-166961; Santa Cruz Biotechnology, Inc.) at 4°C overnight. The relative intensities of the protein signals were normalized to the intensities of the β-actin (clone AC-15, Sigma-Aldrich) signals, and the band densities were quantified by densitometric analysis using ImageJ software (National Institutes of Health, Bethesda, MD, USA; https:// rsb.info.nih.gov/ij/). Tissue homogenates from C57BL/6 mice livers were used as the positive controls [23]. The sample in the vehicle control group with the lowest sEH/β-actin ratio was used as the reference for comparison, and the data are presented as fold changes.  For 14,15-DHET measurements, all samples and standards were assayed in duplicate; the sample in the vehicle control group with the lowest level was used as the reference for comparison and the data are presented as fold changes.

Statistical Analysis.
For maternal and pregnancy characteristics, numerical data are presented as the mean ± SD or median and interquartile range when data were not normally distributed; categorical variables are presented as a number and rate (%). For western blot analysis, real-time qPCR, and ELISA, data are presented as the mean ± SEM. Data were analyzed and plotted using Prism 7 for Mac OS X, version 5.0d (GraphPad Software, Inc., La Jolla, CA, USA). Differences between two groups were computed with the Mann-Whitney U-test or Student's t-test. Differences among groups of three or more were determined by oneway analysis of variance (ANOVA) followed by Bonferroni's post hoc test. A P value of <0.05 was considered statistically significant.

Characteristics of the Study Population.
The characteristics of the study population are shown in Table 1. Compared to normal pregnant women, women with pregnancies complicated by acute CAM had higher body temperature, white blood cell counts, proportions of segmented neutrophils, and baseline fetal heart rate. Furthermore, microbial growth in the amniotic fluid during cesarean delivery was found in 7 of the 16 (44%) patients with acute CAM.

Localization of sEH in the Fetal Membranes from Normal
Pregnant Women and Women with Acute CAM. In normalterm pregnancy, immunostaining of sEH was observed in the amniotic epithelium (Figure 1(g)). The immunostaining of sEH was generally more intense, both in the amnion and choriodecidua, in women with acute CAM than in normal pregnant women (Figure 1(h)). In parallel with the increased immunostaining of sEH in fetal membranes  Figure 3: Immunoreactivity for sEH in villous tissues from normal pregnant women and women with pregnancies complicated by acute CAM. Immunostaining of CK7 delineates the localization of cytotrophoblasts and syncytiotrophoblasts (a, b). Compared to normal pregnancies, villous tissues from women with acute CAM had increased infiltration of monocytes/macrophages, as demonstrated by positive anti-CD68 immunostaining, within the stroma (d, e). In normal pregnancies, immunostaining of sEH was found mainly in the trophoblast layer and some stromal cells (g). The immunostaining of sEH was generally increased in women with acute CAM compared to normal pregnant women (h). In addition, strong sEH immunostaining was noted in the villous endothelium in women with acute CAM (arrows). There was essentially no staining at all on the negative controls when the primary antibody was substituted with nonimmune IgG (c, f, and i). Scale bar = 50 μm.
from pregnancies complicated by acute CAM, there was also increased infiltration of monocytes/macrophages, as demonstrated by positive anti-CD68 immunostaining, within the amnion and choriodecidua (Figures 1(d) and 1(e)).

Increased sEH Protein and mRNA Levels in Fetal
Membranes from Women with Pregnancies Complicated by Acute CAM. Compared with normal pregnant women, women with pregnancies complicated by acute CAM had significantly higher levels of sEH protein and mRNA in fetal membranes ( Figure 2).

Localization of sEH in the Villous Tissues from Normal
Pregnant Women and Women with Acute CAM. As shown in Figure 3(g), immunoreactivity for sEH was found in the trophoblast layer of villous tissues obtained from normal pregnant women. A few stromal cells also reacted to the anti-sEH antibody. In villous samples from pregnancies complicated by acute CAM, the immunostaining of sEH was generally more intense-in particular in the trophoblast layer and villous endothelium-than in those from normal pregnancies lacking the symptoms and signs of CAM (Figure 3(h)). This change was associated with increased infiltration of monocytes/macrophages into the stroma of the villi (Figures 3(d) and 3(e)).

Increased sEH Protein and mRNA Levels in Villous
Tissues from Women with Pregnancies Complicated by Acute CAM. Similarly, the villous samples of women with pregnancies complicated by acute CAM had significantly higher levels of sEH protein and mRNA than those of normal pregnant women (Figure 4).

LPS Led to Increased sEH mRNA and Protein Levels in
Fetal Membrane and Villous Explants. The most wellrecognized predisposing factor to acute CAM is ascending infection by microorganisms from the lower genital tract. We therefore studied the effect of bacterial endotoxin, LPS, on the expression of sEH in fetal membrane and villous explant cultures. Compared to the vehicle controls, administration of LPS in the culture medium significantly increased the levels of sEH mRNA and protein in the amnion and choriodecidua with the two-chamber culture system ( Figures 5(a), 5(b), 5(d), and 5(e) and Supplementary  Figures 3b and 3c). Similar effects of LPS on the expression of sEH were also observed in villous explant culture experiments (Figures 5(c) and 5(f) and Supplementary  Figure 3a).

Administration of AUDA Reduced the Tissue Levels of 14,15-DHET in Fetal Membrane and Villous Explants
Treated with LPS. 14,15-DHET is one of the major DHETs produced by sEH in human placentas; therefore, measurement of 14,15-DHET can be regarded as an indicator of the enzymatic activity of sEH. To study the changes of sEH activity in response to LPS stimulation with or without AUDA, we measured the tissue levels of 14,15-DHET in the homogenates of fetal membrane and villous explants, respectively.

Discussion
Our results demonstrate that (1) women with pregnancies complicated by acute CAM had more intense immunostaining of sEH and higher levels of sEH mRNA and protein in fetal membranes and villous tissues than women with normal-term pregnancies without CAM; results suggest that sEH participates in the inflammatory changes in human gestational tissues in pregnancies complicated by acute CAM. sEH expression has been detected in the gravid uterus and placenta including fetal membranes [18,19,25]; however, its role in these tissues remains elusive. Using myometrial strips obtained from normal pregnant women experiencing cesarean deliveries, Corriveau et al. found that administration of EETs or AUDA to the tissue bath significantly inhibited spontaneous myometrial contractile activities in vitro [25]. They further showed that the levels of sEH were greater in the myometrium of LPS-treated pregnant rats than in pregnant rats without LPS stimulation and that myometrial contractile activity induced by LPS was reduced upon administration of EETs or AUDA [18]. In contrast, women with preeclampsia have higher plasma and placental levels of EETs than normal pregnant women [19,26,27]. Furthermore, the expression of sEH was lower in preeclamptic placentas compared to normal placentas [19]. The current study expands upon previous findings of the presence of sEH in human gestational tissues and provides new evidence of the role of placental sEH in pregnancy complications including acute CAM.
The mechanisms underlying the regulation of sEH expression in human gestational tissues in pregnancies complicated by CAM and in response to LPS stimulation remain unclear. Our recent study shows that NF-κB signaling is a possible pathway [23]. The sEH gene promoter region contains recognition sites for a number of transcription factors including NF-κB [28]. NF-κB is a key regulator of proinflammatory cytokine production in response to infection [29]. Indeed, increased NF-κB activity has been detected in the placentas of pregnancies complicated by acute CAM compared to those without acute CAM [30]. Furthermore, activation of NF-κB has been observed in placentas perfused with LPS ex vivo [31] and in villous and fetal membrane explants treated with LPS in vitro [32].
In addition to its effect on the induction of sEH, we found that LPS treatment simultaneously caused increased levels of 14,15-DHET in the tissue homogenates of fetal membrane and villous explants, suggesting an increased activity of sEH in these tissues. We also observed that the administration of AUDA reduced the levels of 14,15-DHET induced by LPS, confirming the inhibitory effect of AUDA on sEH activity. We further demonstrated that the administration of AUDA reduced the LPS-stimulated production of IL-1β and IL-6 in fetal membrane and villous explants and in pregnant mice. These findings suggest an association between the pharmacological inhibition of sEH and the reduced production of cytokines in response to LPS stimulation, though the mechanisms regulating these changes are not clear. Previous studies on other organ systems have shown that the administration of EETs or AUDA attenuated the phosphorylation of mitogen-activated protein (MAP) kinases such as c-Jun N-terminal kinases (JNK) and p38 [23,33,34]. MAP kinases have been reported to participate in the inflammatory response to LPS stimulation in the cells or tissues obtained from pregnant women including amnion cells [35], decidual cells [36], cytotrophoblasts from term placenta [37], and choriodecidual explants [38]. The activation of MAP kinase signaling contributes to the activation of several transcriptional factors including NF-κB, resulting in the upregulation of the genes of many proinflammatory proteins such as cytokines and chemokines [39]. The activation of NF-κB can also increase the expression of COX and LOX, leading to the production of PGs and LTB 4 [29]. Inflammatory cytokines, chemokines, PGs, and LTB 4 have been reported to play essential roles in CAM-related preterm labor and birth [40]. Therefore, by inhibiting the activity of sEH-thus stabilizing endogenous EETs-AUDA likely exerts its antiinflammatory effect via the attenuation of phosphorylation of MAP kinases and subsequent activation of NF-κB in human gestational tissues with bacterial infection. Further studies are needed to verify our assumption and such results would be of clinical relevance, forming the basis of new therapeutic strategies in addition to antibiotic therapy in treating women with CAM.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
All the authors declare that they have no conflicts of interest.
LPS at a concentration of 1 μg/mL, based on one-way ANOVA followed by Bonferroni's post hoc test. Supplementary Figure 2: after confirming the effect of LPS on the changes in sEH, we next studied the effect of AUDA on LPS-induced changes in levels of 14,15-DHET, IL-1β, and IL-6 in villous and fetal membrane explants. To determine the working concentration of AUDA, villous explants were incubated with LPS at a concentration of 10 μg/mL and various concentrations of AUDA (0, 1, 10, and 100 μM) dissolved in DMSO for 24 h and the levels of IL-1β and IL-6 in the medium were measured. As shown, the administration of AUDA at a concentration of 10 or 100 μM significantly reduced the levels of IL-1β and IL-6 in the medium induced by LPS, with no difference in the effect between the two concentrations. Therefore, AUDA at a concentration of 10 μM was used to study its effect on LPS-induced changes in 14,15-DHET, IL-1β, and IL-6 in villous and fetal membrane explants in the subsequent experiments. Data are presented as mean ± SEM, n = 4 for each group. * * * P < 0:001 compared with the vehicle control; ### P < 0:001 compared with explants treated with LPS (10 μg/mL) alone, based on one-way ANOVA followed by Bonferroni's post hoc test. Supplementary Figure 3: representative blots of whole gel showing that the administration of LPS at a concentration of 10 μg/mL to the culture medium led to a significant increase in the levels of sEH protein in villous explants (a), amnion (b), and choriodecidua (c) with the twochamber culture system (n = 4 for each group). C57BL/6 mouse liver homogenate was used as the positive control for sEH (a). β-Actin was used to normalize loading variability. Molecular weight markers are shown at the left side of the blots. (Supplementary Materials)