Endothelial CYP epoxygenase overexpression and soluble epoxide hydrolase disruption attenuate acute vascular inflammatory responses in mice

Abstract

Cytochrome P-450 (CYP)-derived epoxyeicosatrienoic acids (EETs) possess potent anti-inflammatory effects in vitro. However, the effect of increased CYP-mediated EET biosynthesis and decreased soluble epoxide hydrolase (sEH, Ephx2)-mediated EET hydrolysis on vascular inflammation in vivo has not been rigorously investigated. Consequently, we characterized acute vascular inflammatory responses to endotoxin in transgenic mice with endothelial expression of the human CYP2J2 and CYP2C8 epoxygenases and mice with targeted disruption of Ephx2. Compared to wild-type controls, CYP2J2 transgenic, CYP2C8 transgenic, and Ephx2^−/− mice each exhibited a significant attenuation of endotoxin-induced activation of nuclear factor (NF)-κB signaling, cellular adhesion molecule, chemokine and cytokine expression, and neutrophil infiltration in lung in vivo. Furthermore, attenuation of endotoxin-induced NF-κB activation and cellular adhesion molecule and chemokine expression was observed in primary pulmonary endothelial cells isolated from CYP2J2 and CYP2C8 transgenic mice. This attenuation was inhibited by a putative EET receptor antagonist and CYP epoxygenase inhibitor, directly implicating CYP epoxygenase-derived EETs with the observed anti-inflammatory phenotype. Collectively, these data demonstrate that potentiation of the CYP epoxygenase pathway by either increased endothelial EET biosynthesis or globally decreased EET hydrolysis attenuates NF-κB-dependent vascular inflammatory responses in vivo and may serve as a viable anti-inflammatory therapeutic strategy.—Deng, Y., Edin, M. L., Theken, K. N., Schuck, R. N., Flake, G. P., Kannon, M. A., DeGraff, L. M., Lih, F. B., Foley, J., Bradbury, J. A., Graves, J. P., Tomer, K. B., Falck, J. R., Zeldin, D. C., Lee, C. R. Endothelial CYP epoxygenase overexpression and soluble epoxide hydrolase disruption attenuate acute vascular inflammatory responses in mice.

Inflammation is a host reaction characterized by endothelial activation, leukocyte–endothelial adhesion, and the subsequent transmigration of leukocytes into tissue (1). It is well established that nuclear factor-κB (NF-κB) is a central mediator of this coordinated process in vivo through transcriptional activation of cellular adhesion molecule (CAM), chemokine, and cytokine expression (2–4). Consequently, NF-κB activation is integral to the pathogenesis of numerous acute and chronic inflammatory diseases in humans, such as sepsis, asthma, and atherosclerosis (4–7).

Cytochrome P-450 (CYP) epoxygenases from the CYP2J and CYP2C subfamilies catalyze the oxidative metabolism of arachidonic acid to epoxyeicosatrienoic acids (EETs) in various cell types, including endothelial cells (8), and possess potent vasodilatory (9), proangiogenic (10), antiapoptotic (11), and postischemia protective effects (12) in the vasculature. Soluble epoxide hydrolase (sEH, Ephx2) rapidly hydrolyzes EETs to their corresponding dihydroxyeicosatrienoic acid (DHET) metabolites, which are less biologically active than EETs (13). Pharmacological inhibition of sEH increases cellular and circulating EET levels, potentiates the biological effects of EETs in preclinical models, and is currently under clinical development as a novel therapeutic strategy for hypertension (14).

Recent studies have demonstrated that CYP-derived EETs also possess potent anti-inflammatory effects (8). Most notably, EET administration and CYP2J overexpression significantly attenuated cytokine-induced NF-κB activation and CAM expression in cultured endothelial cells, and exogenous EET administration attenuated leukocyte adhesion to isolated-perfused murine arterioles (15). More recently, induction of circulating cytokine levels and mortality by endotoxin was significantly reduced in mice with targeted disruption of Ephx2 (16) and wild-type mice treated with an sEH inhibitor (17), demonstrating the systemic anti-inflammatory effects elicited by increased circulating EET levels. However, the functional effect of CYP-derived EETs on the regulation of local, NF-κB-dependent inflammatory responses in vivo remains unclear.

Consequently, we developed transgenic mice with endothelial expression of the primary CYP epoxygenases responsible for EET biosynthesis in humans, CYP2J2 and CYP2C8 (18), and characterized acute inflammatory responses to endotoxin, a well-established activator of NF-κB signaling, endothelial activation and leukocyte trafficking in preclinical models and humans (3, 6, 19–21). Mice with targeted disruption of Ephx2 were also included in these experiments since the relative effect of increased CYP-mediated EET biosynthesis and decreased sEH-mediated EET hydrolysis on inflammation has not been investigated to date. Our findings demonstrate that increased endothelial CYP2J2- and CYP2C8-mediated EET biosynthesis and globally decreased sEH-mediated EET hydrolysis similarly attenuate acute, NF-κB-dependent vascular inflammatory responses in vivo.

MATERIALS AND METHODS

Chemicals

Chemicals were purchased from Sigma (St. Louis, MO, USA) unless otherwise noted.

Animals

Transgenic (Tr) mice that express either human CYP2J2 (Tie2-CYP2J2-Tr) or CYP2C8 (Tie2-CYP2C8-Tr) in endothelial cells were developed on a pure C57BL/6 background, as described previously (18). Briefly, the human CYP2J2 (GenBank accession number NM_000775) and CYP2C8 (NM_000770) cDNA sequences were subcloned downstream of the murine Tie2 promoter (2.1 kb) and upstream of the Tie2 full enhancer (10 kb) (generously provided by Dr. Tom Sato, University of Texas Southwestern Medical Center, Dallas, TX, USA) to drive endothelial expression of human CYP2J2 and CYP2C8 in mice. A colony of mice with targeted disruption of Ephx2 (Ephx2^−/−) was rederived at the National Institute of Environmental Health Sciences (NIEHS) and backcrossed onto a C57BL/6 genetic background for >10 generations (22, 23). Mice were genotyped by PCR and bred as described previously (18, 23), had free access to food and water, and were housed in controlled conditions for temperature and humidity using a 12-h light-dark cycle. All studies utilized Tr mice from multiple founder lines and wild-type littermates as controls, were completed in accordance with the U.S. National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals, and were approved by the Institutional Animal Care and Use Committee at the University of North Carolina at Chapel Hill and National Institutes of Health/NIEHS.

Primary endothelial cell isolation

Lungs from Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, Ephx2^−/−, and wild-type littermates (age 8–10 wk) were harvested for isolation of primary endothelial cells, as described previously (18, 24). Briefly, endothelial cells were isolated from lung homogenates by magnetic separation using Dynabeads (Invitrogen, Carlsbad, CA, USA) coated with anti-PECAM (BD Biosciences, San Jose, CA, USA), and then plated into flasks coated with 20 μg/ml type I collagen (Collaborative Biomedical Sciences, Bedford MA, USA) in DMEM containing 20% FBS, 100 μg/ml heparin, 100 μg/ml penicillin/streptomycin, and 100 μg/ml endothelial cell growth supplement (Biomedical Technologies, Stoughton, MA) with nonessential amino acids, 2 mM l-glutamine, 2 mM sodium pyruvate and 25 mM HEPES. After ∼5–9 d in culture at 37°C in 5% CO₂ and 95% air, confluent monolayers were trypsinized and further purified by incubation with anti-ICAM (BD Biosciences)-coated Dynabeads and grown to confluence.

For basal characterization studies, cells were plated on 60-mm tissue culture dishes in serum-free medium with a Ca²⁺ ionophore (A23187, 10 μM × 15 min) to stimulate arachidonic acid release from the cell membrane. For inflammation studies, cells were plated into 6-well dishes, cultured for 24 h, and then treated with Escherichia coli LPS (serotype O111:B4; 1, 10, or 100 ng/ml) or vehicle (endotoxin-free PBS) for 4 h. Thirty minutes prior to LPS administration, a subset of cells was incubated with the putative EET receptor antagonist 14,15-epoxyeicosa-5-(Z)-enoic acid (14,15-EEZE, 10 μM) (18, 25), the selective CYP epoxygenase inhibitor N-(methylsulfonyl)-2-(2-propynyloxy)-benzenehexanamide (MS-PPOH, 10 μM; Cayman Chemical, Ann Arbor, MI, USA) (26) or vehicle (0.01% ethanol). Cells and media were harvested and stored at −80°C for RNA/protein isolation and eicosanoid extraction.

Induction of endotoxemia in mice

Adult male and female Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, Ephx2^−/−, and wild-type littermate control mice weighing 22–35 g received a single dose of Escherichia coli LPS (serotype O111:B4, 10⁶ EU/mg; 40 mg/kg) or endotoxin-free PBS by intraperitoneal injection. Pilot studies demonstrated peak induction of proinflammatory gene expression 3 h after LPS dosing. Consequently, all mice were euthanized by CO₂ inhalation 3 h after dosing. Blood was collected by cardiac puncture, and plasma was separated by centrifugation. Lung tissue was snap-frozen in liquid nitrogen and stored at −80°C for RNA/protein isolation, or fixed in 4% paraformaldehyde and embedded in paraffin for histology.

Real-time quantitative RT-PCR

Total RNA was isolated from lysed primary lung endothelial cells (∼10⁶ cells/isolation) and lung homogenates using the RNeasy Miniprep kit (Qiagen, Valencia, CA, USA) per the manufacturer's instructions. Total RNA was reverse transcribed to cDNA using the ABI high-capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA, USA) with a reaction temperature of 25°C for 10 min, and then 37°C for 120 min. Quantitative PCR was performed on an ABI 7300 real-time PCR System (Applied Biosystems). Expression of human CYP2J2 (Hs00356031_m1) and CYP2C8 (Hs00426387_m1), and murine IL6 (Mm00446190_m1), IL1B (Mm00434228_m1), CCL2 (Mm00441242_m1), CXCL5 (Mm01354316_g1), SELE (Mm00441278_m1), and GAPDH (Mm99999915_g1) were quantified using Taqman Assays on Demand (Applied Biosystems). Each amplification reaction was performed in triplicate and carried out in a 20-μl reaction volume with 50 ng cDNA, 20× Assay on Demand Gene Expression kit (1 μl) and 2× Taqman Universal PCR Master Mix (10 μl), and was completed as follows: 2 min at 50°C (1 cycle); 10 min at 95°C (1 cycle); 15 s at 95°C, and 60 s at 60°C (40 cycles). Gene expression was normalized to GAPDH and expressed relative to saline-treated wild-type controls using the 2^−ΔΔCt method (27).

Immunoblotting

Microsomal fractions from whole lung were prepared by differential centrifugation at 4°C, as described previously (28). Lung tissue or primary endothelial cells were homogenized in ice-cold lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Triton X, 1 mM NaF, and 0.25% Na deoxycholate and protease inhibitors, and the S9 fraction was separated by centrifugation. Protein concentrations were quantified by Bradford's method (29).

Equal amounts of primary endothelial cell lysates (7–15 μg), lung microsomes (70 μg), or lung homogenates (50 μg) were separated by 10% NuPAGE Bis-Tris gels, transferred to nitrocellulose membranes (Invitrogen), and blocked in 5% nonfat milk in Tris-buffered saline (TBS). For primary endothelial cell lysates, lung lysates, and lung microsomes isolated from untreated mice, membranes were incubated with 1:1000 dilutions of the anti-CYP2J2pep1, anti-CYP2J2pep3, anti-CYP2C8 (kindly provided by Dr. Joyce Goldstein, NIEHS, Research Triangle Park, NC, USA), anti-sEH (sc22344; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and anti-β-actin (4967; Cell Signaling Technology, Danvers, MA, USA) antibodies, as described previously (18). Recombinant CYP2J2 and CYP2C8 proteins were included as positive controls (30, 31). Anti-CYP2J2pep1 (aa 103–117, immunospecific for human CYP2J2; ref. 32) and anti-CYP2J2pep3 (aa 356–379, cross-reactivity with murine CYP2J isoforms; ref. 30) are purified rabbit polyclonal antibodies raised against CYP2J2-specific peptides. Anti-CYP2C8 is a purified rabbit polyclonal antibody raised against recombinant human CYP2C8; however, cross-reactivity with murine CYP2C isoforms exists (33). For primary endothelial cell lysates and lung lysates isolated from the LPS experiments, membranes were incubated at 4°C overnight with a 1:1000 dilution of the anti-IκBα (4812), anti-phospho-IκBα (2859), and anti-GAPDH (2118) antibodies (Cell Signaling Technology) in 1% BSA in TBS with 0.05% Tween 20. Membranes were then washed and incubated with a 1:2000 dilution of horseradish peroxidase-conjugated bovine anti-rabbit secondary antibody (Santa Cruz Biotechnology). Immunoreactive bands were detected by chemiluminescence using the SuperSignal West Dura substrate (Thermo Scientific, Rockford, IL, USA) and visualized with the VersaDoc Imaging system (Bio-Rad, Hercules, CA, USA). Intensity of each immunoreactive band for IκBα and phospho-IκBα was quantified by densitometry using Quantity One software (Bio-Rad), normalized to GAPDH, and expressed relative to the saline-treated wild-type control group.

ELISA

MCP-1 and E-selectin protein levels were quantified in primary endothelial cell lysates and lung homogenates using the Quantikine mouse CCL2/JE/MCP-1 and E-selectin immunoassay kits, respectively, according to the manufacturer's instructions (R&D Systems, Minneapolis, MN, USA). MCP-1 and E-selectin protein levels were calculated as picogram per milligram of total protein, and expressed relative to the saline-treated wild-type control group.

Quantification of CYP-derived eicosanoids

Epoxy (EET) and dihydroxy (DHET) metabolites of arachidonic acid were extracted from primary endothelial cell culture medium and plasma by solid-phase extraction and quantified by HPLC-MS/MS, as described previously (18, 23, 34). Medium concentrations were normalized to cell density. In particular, the 11,12- and 14,15-regioisomers were quantified since these are the primary metabolic products synthesized by human CYP2J2 and CYP2C8 (30, 35).

Histology and immunohistochemistry

Lung tissue was fixed with 4% paraformaldehyde, processed, embedded in paraffin, and cut into 5- to 6-μm sections. For baseline characterization, lung sections underwent immunohistochemical staining with the anti-CYP2J2pep1 and anti-CYP2C8 antibodies, as described previously (18). For the LPS experiments, lung sections were stained with hematoxylin and eosin (H&E), and were also immunohistochemically stained for myeloperoxidase (MPO). Briefly, slides were placed in 3% H₂O₂ for 15 min to quench endogenous peroxide, and then boiled in 0.01 M sodium citrate buffer (pH 6.0) for 20 min to retrieve antigen for immunohistochemistry. Sections were incubated with rabbit polyclonal anti-MPO (1:1500, Dako Corp., Carpinteria, CA, USA) for 30 min at room temperature, and antibody binding was detected using Vectastain Elite ABC Kit (Vector Laboratories, Burlingame, CA, USA). Visualization was performed with 3,3′-diaminobenzidine (Vector Laboratories) as chromogen. Slides were counterstained with hematoxylin, dehydrated, and mounted. Stained sections were observed under a light microscope, and digital images were captured.

Histopathological assessment of the LPS response was conducted by a pathologist who was masked to genotype and treatment group assignments using two methods: H&E staining and MPO immunohistochemical staining. First, the H&E-stained sections were scored for the presence of neutrophilic margination in vessels (0–1), perivascular neutrophils (0–4), peribronchial neutrophils (0–4), and intra-alveolar neutrophils (0–4), as described previously (36), and a sum total was calculated for each slide. Second, MPO-stained slides were evaluated by scoring (0–4) for the presence of neutrophils within the alveolar walls, which is not possible with H&E staining since neutrophils are not readily distinguished within the alveolar walls.

Myeloperoxidase activity

MPO is a leukocyte-derived heme oxidase primarily carried by polymorphonuclear neutrophils (37). Since detection of MPO enzymatic activity in lung parenchyma predominantly reflects the presence of neutrophils (38), MPO functional activity was quantified to further assess neutrophil infiltration. Lung (30 mg) was homogenized in 20 mM phosphate buffer saline with 10 mM N-ethylmaleimide (pH 7.4), and centrifuged at 10,000 g for 15 min at 4°C. The pellet was resuspended in 50 mM PBS (pH 6.0) containing 0.5% hexadecyltrimethylammonium bromide, sonicated for 30 s, and centrifuged at 8000 g for 20 min at 4°C. MPO activity in the supernatant was quantified using the EnzChek MPO activity assay kit (Invitrogen), according to the manufacturer's instructions. Fluorescence was quantified on a Fluoroskan Ascent FL luminometer (Thermo Scientific) at excitation/emission wavelengths of 480/530 nm, and expressed relative to the saline-treated wild-type control group.

Statistical analysis

Data were expressed as means ± se, normalized to the saline-treated wild-type (control) group, and compared statistically across genotype groups using a 1-way ANOVA followed by a post hoc Student's t test. Statistical analysis was performed using SAS-JMP 6.0 software (SAS Institute, Cary, NC, USA). P < 0.05 was considered statistically significant.

RESULTS

Baseline characterization of Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice

Tie2-CYP2J2-Tr and Tie2-CYP2C8-Tr mice exhibited significantly higher CYP2J2 and CYP2C8 transcript levels, respectively, in RNA isolated from primary lung endothelial cells (Fig. 1A) and lung tissue homogenates (Fig. 1C, F), and higher CYP2J2 and CYP2C8 protein expression in primary lung endothelial cell lysates (Fig. 1A) and lung microsomal fractions (Fig. 1D, G) compared to wild-type littermates. Immunohistochemical staining of lung tissue sections demonstrated endothelial CYP2J2 and CYP2C8 expression in Tie2-CYP2J2-Tr (Fig. 1E) and Tie2-CYP2C8-Tr mice (Fig. 1H), respectively, while wild-type littermates exhibited minimal to no expression. Primary lung endothelial cells isolated from Tie2-CYP2J2-Tr and Tie2-CYP2C8-Tr mice released ∼2-fold higher concentrations of 11,12- and 14,15-EET and DHET into culture medium compared to cells isolated from wild-type controls (Fig. 1B), demonstrating higher endothelial CYP epoxygenase metabolic activity in Tr mice.

Figure 1.

Baseline characterization of Tie2-CYP2J2-Tr and Tie2-CYP2C8-Tr mice. A) Human CYP2J2 and CYP2C8 mRNA levels (assessed by quantitative RT-PCR) and protein levels (assessed by immunoblotting) were abundant in primary lung endothelial cells (ECs) isolated from Tr mice, but not wild-type (WT) littermates. B) After stimulation with A23187, concentrations of 11,12- and 14,15-EET and DHET (the stable EET metabolite), and the regioisomer sum total, released into primary lung EC medium are significantly higher in Tr mice compared to WT littermates (n=3 isolations/genotype group). *P < 0.05 vs. WT. C, F) Human CYP2J2 (C) and CYP2C8 (F) mRNA levels in lung homogenates are significantly higher in Tr mice from multiple founder lines compared to WT littermates (n=11–12 mice/genotype group). *P < 0.001 vs. WT. D, G) Representative immunoblot of microsomal fractions isolated from whole-lung homogenates demonstrate higher CYP2J2 (D) and CYP2C8 (G) protein expression in Tr mice from multiple founder lines (18) compared to WT littermates. Densitometry analysis demonstrated 1.4 ± 0.1- and 1.6 ± 0.2-fold higher immunoreactivity to anti-CYP2J2pep3 and anti-CYP2C8, respectively, in Tr compared to WT mice. Recombinant CYP2J2 and CYP2C8 protein were included as positive controls and exhibited a slightly higher molecular mass compared to the endogenous protein, consistent with prior reports (30, 31). E, H) CYP2J2 (E) and CYP2C8 (H) immunostaining was completed using the anti-CYP2J2pep1 and anti-CYP2C8 antibodies, respectively. Images are from representative lung sections of Tr and WT littermate control mice. Arrows identify endothelial cell staining in Tr mice, but not WT littermates. No immunostaining was observed with normal rabbit serum (not shown).

Absence of sEH protein was confirmed in primary lung endothelial cells (Fig. 2A) and lung tissue homogenates (Fig. 2B) from Ephx2^−/− mice by immunoblotting. Plasma 11,12-EET:DHET (2.39±0.67 vs. 0.68±0.19, P=0.023) and 14,15-EET:DHET (2.09±0.28 vs. 0.19±0.04, P<0.001) epoxide to diol ratios were significantly higher in Ephx2^−/− mice compared to wild-type littermate controls, respectively, consistent with sEH disruption (16, 23). Consequently, plasma 11,12- and 14,15-EET concentrations were ∼3.5-fold higher in Ephx2^−/− mice (Fig. 2C). We previously reported an ∼1.5-fold elevation in circulating EET concentrations in Tie2-CYP2J2-Tr and Tie2-CYP2C8-Tr mice compared to wild-type littermate controls (18).

Figure 2.

Baseline characterization of Ephx2^−/− mice. A, B) Expression of murine sEH protein (assessed by immunoblotting) in primary lung endothelial cells (ECs) (A) and whole lung homogenates (B) was absent in Ephx2^−/− mice, but not wild-type (WT) littermates. C) Plasma concentrations of 11,12- and 14,15-EET, and the regioisomer sum total, are significantly higher in Ephx2^−/− mice compared to WT littermates (n=7–8/genotype group). *P < 0.01 vs. WT.

Primary endothelial cell activation in vitro

In primary endothelial cells isolated from wild-type mice, LPS activated E-selectin and MCP-1 mRNA expression in a concentration-dependent manner; however, this up-regulation was significantly attenuated in primary endothelial cells isolated from Tie2-CYP2J2-Tr and Tie2-CYP2C8-Tr mice (Fig. 3A, B). In a subsequent experiment, administration of the putative EET receptor antagonist 14,15-EEZE or the selective CYP epoxygenase inhibitor MS-PPOH significantly weakened the attenuation of E-selectin and MCP-1 mRNA expression observed in endothelial cells isolated from Tie2-CYP2J2-Tr mice (Fig. 3C, D). In contrast, administration of 14,15-EEZE or MS-PPOH did not affect E-selectin or MCP-1 expression in endothelial cells isolated from wild-type littermate controls.

Figure 3.

Endothelial activation by LPS in vitro. A, B) In primary lung endothelial cells (ECs), E-selectin (A) and MCP-1 mRNA (B) levels were increased by LPS (1, 10, 100 ng/ml) in a dose-dependent manner and significantly attenuated in Tie2-CYP2J2-Tr and Tie2-CYP2C8-Tr mice compared to wild-type (WT) littermate controls. C–F) E-selectin (C, E) and MCP-1 (D, F) mRNA and protein levels were also quantified in primary ECs isolated from Tie2-CYP2J2-Tr mice and WT littermate controls (n=3/group) after stimulation with LPS (100 ng/ml) in the absence (vehicle) and presence of 14,15-EEZE (10 μM) or MS-PPOH (10 μM). mRNA data (A–D) were normalized to GAPDH using the 2^−ΔCt method. Protein data (E, F) were expressed relative to the saline-treated WT control group. G) Representative immunoblot evaluating phosphorylated IκBα immunoreactivity in EC lysates from the same experiment. H) Densitometry analysis of phosphorylated IκBα immunoreactivity was normalized to GAPDH and expressed relative to the saline-treated WT control group (n=3/group). All data are presented as means ± se. *P < 0.05 vs. WT within same treatment group; ^#P < 0.05 vs. vehicle within same genotype group.

Induction of MCP-1 protein expression was also significantly attenuated in endothelial cells isolated from Tie2-CYP2J2-Tr mice (Fig. 3F, P<0.05 vs. wild-type), which was weakened by administration of 14,15-EEZE (P=0.102) and MS-PPOH (P<0.05). Similar results were observed with E-selectin protein expression, although these differences did not attain statistical significance (Fig. 3E). In addition, the LPS-mediated increase of phosphorylated IκBα levels (a marker of NF-κB activation) in endothelial cells isolated from wild-type mice was significantly attenuated in endothelial cells isolated from Tie2-CYP2J2-Tr mice, and 14,15-EEZE and MS-PPOH each significantly inhibited the attenuation of IκBα phosphorylation in Tie2-CYP2J2-Tr cells (Fig. 3G, H).

Induction of the acute inflammatory response in vivo

In vivo, administration of LPS significantly up-regulated proinflammatory cytokine (IL-6 and IL-1β, Fig. 4A), chemokine (MCP-1 and ENA-78, Fig. 4B), and CAM (E-selectin, Fig. 4C) mRNA levels in lung in wild-type mice relative to saline-treated controls. Compared to wild-type littermates, the LPS-mediated induction of each proinflammatory mediator was significantly attenuated in Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice (Fig. 4; P<0.05 vs. LPS-treated wild-type mice), although the observed attenuation of ENA-78 mRNA levels in Tie2-CYP2C8-Tr mice was not statistically significant (P=0.090). No differences across founder lines or gender were observed. Similarly, the LPS-mediated induction of E-selectin (Fig. 5A) and MCP-1 (Fig. 5B) protein expression was significantly attenuated in Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr and Ephx2^−/− mice compared to wild-type littermates (P<0.05), although the observed attenuation of MCP-1 in Tie2-CYP2C8-Tr mice was not statistically significant (P=0.067).

Figure 4.

Induction of cytokine, chemokine, and cellular adhesion molecule mRNA expression in vivo. IL-6 and IL-1β (A), MCP-1 and ENA-78 (B), and E-selectin (C) mRNA levels were quantified in lung homogenates 3 h after LPS or saline administration by quantitative RT-PCR. Expression was significantly attenuated in LPS-treated Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice (n=15–18/group) compared to wild-type (WT) littermates (n=35). Data are presented as mean ± se-fold change in expression, relative to the WT-saline control group (n=21), using the 2^−ΔΔCt method. No differences were observed across genotype groups in saline-treated mice (n=4–6/group, data not shown). *P < 0.05 vs. LPS-treated WT group.

Figure 5.

Induction of proinflammatory mediator protein expression and nuclear factor-κB activation in vivo. A, B) E-selectin (A) and MCP-1 (B) protein levels were quantified in lung homogenates 3 h after LPS or saline administration by ELISA. Expression was significantly attenuated in LPS-treated Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice compared to wild-type (WT) littermates (n=8–10/group). C) Representative immunoblot evaluating IκBα, phosphorylated IκBα, and GAPDH immunoreactivity in lung homogenates 3 h after saline or LPS administration. D, E) Densitometry analysis of phosphorylated IκBα immunoreactivity was normalized to GAPDH (D) and IκBα (E). Each ratio was expressed relative to saline-treated WT controls (n=10), and was lower in Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice compared to WT littermates (n=6–9/group). Data are presented as means ± se. *P < 0.05 vs. LPS-treated WT group.

Compared to saline-treated controls, administration of LPS markedly activated NF-κB in wild-type mice in vivo, which was reflected by significantly higher abundance of phosphorylated IκBα protein levels in lung relative to GAPDH and total IκBα (Fig. 5C–E). LPS-mediated IκBα phosphorylation was significantly lower in Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice compared to wild-type littermates (P<0.05), except in Tie2-CYP2C8-Tr mice when calculated relative to total IκBα (Fig. 5E, P=0.062), demonstrating a significant attenuation of NF-κB activation in the presence of CYP epoxygenase pathway potentiation in vivo.

Neutrophil infiltration

No histopathological evidence of inflammation was observed in the absence of LPS (Supplemental Fig. S1A and Fig. 6D), such that the H&E and MPO staining methods each yielded an inflammatory score of zero in all saline-treated mice (Fig. 6A, B). Because of the relatively early time point used in this experiment, the inflammatory response detected by H&E staining was limited primarily to neutrophilic margination within the vessels (Supplemental Fig. S1); however, perivascular neutrophils were also observed in 3 of the 7 LPS-treated wild-type mice. In contrast, no perivascular neutrophils were detected in Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice. Moreover, Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice showed a marked and statistically significant attenuation of neutrophil infiltration compared to LPS-challenged wild-type littermate controls (Fig. 6A; P<0.05 vs. wild-type).

Figure 6.

Histopathological and functional assessment of neutrophil infiltration in vivo. A, B) Semiquantitative score evaluating neutrophil infiltration on H&E-stained (A) and myeloperoxidase (MPO)-immunostained (B) lung sections. All saline-treated wild-type (WT) mice had a score of 0 (n=5). A lower score was observed in LPS-treated Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice compared to WT littermates (n=5–7/group). C) MPO activity in lung homogenates, expressed relative to saline-treated WT controls (n=12), was significantly lower in LPS-treated Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice compared to WT littermates (n=10–14/group). Data are presented as means ± se. *P < 0.05 vs. LPS-treated WT group. D–H) Representative images of MPO-immunostained lung sections (×20) in saline-treated WT (D) and LPS-treated WT (E), Tie2-CYP2J2-Tr (F), Tie2-CYP2C8-Tr (G), and Ephx2^−/− (H) mice. Arrows indicate MPO staining. Although the saline-treated WT image demonstrates margination of 3 neutrophils within the vessel, this image was selected to confirm the presence of positive immunostaining on the slide. The vast majority of vessels in saline-treated WT mice demonstrated no evidence of neutrophil margination. Representative images of H&E-stained sections are provided in Supplemental Fig. S1.

Although neutrophils were not readily distinguished within the alveolar walls with H&E staining, immunohistochemical staining demonstrated alveolar wall infiltration of MPO-positive cells after LPS administration (Fig. 6E–H). Alveolar wall infiltration of MPO-positive cells was significantly attenuated in Tie2-CYP2J2-Tr and Ephx2^−/− mice (Fig. 6B, P<0.05 vs. wild-type). Although Tie2-CYP2C8-Tr mice also exhibited fewer MPO-positive cells, these differences were not statistically significant (Fig. 6B, P=0.094 vs. wild-type). Similarly, relative to saline controls, LPS induced a >2-fold increase in pulmonary MPO functional activity in wild-type mice (Fig. 6C). Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice each exhibited a significantly attenuated induction of MPO activity in lung tissue (P<0.05 vs. LPS-treated wild-type mice).

DISCUSSION

Cytochrome P-450 CYP2J and CYP2C epoxygenases catalyze the biosynthesis of EETs, which possess numerous vascular effects and are rapidly hydrolyzed by sEH (8, 14). Although recent in vitro studies have demonstrated that CYP-derived EETs possess potent anti-inflammatory effects in endothelial cells via inhibition of NF-κB activation (8), the functional contribution of the CYP epoxygenase pathway to the regulation of vascular inflammatory responses in vivo has not been rigorously characterized. This study is the first to demonstrate that increased CYP2J2 and CYP2C8-mediated endothelial EET biosynthesis and decreased sEH-mediated EET hydrolysis similarly attenuate induction of acute, NF-κB-dependent, vascular inflammatory responses in vivo. These findings suggest that potentiation of the CYP epoxygenase pathway may represent a viable anti-inflammatory therapeutic strategy.

Although recent studies have demonstrated that genetic disruption or pharmacological inhibition of sEH elicits anti-inflammatory effects in vivo (14), the direct contribution of CYP-mediated EET biosynthesis to the regulation of NF-κB-dependent vascular inflammation has remained unclear as a result of the lack of selective pharmacological tools conducive to in vivo dosing (8). To characterize the functional contribution of endothelial CYP epoxygenase-derived EETs to the regulation of this fundamental pathophysiological process in vivo, we developed transgenic mice that exhibit constitutive endothelial expression of the primary human CYP epoxygenases, CYP2J2 and CYP2C8, and increased endothelial EET biosynthesis (18), and evaluated the acute inflammatory response following endotoxin administration. Mice with targeted disruption of the murine Ephx2 gene, which have significantly increased EET levels in multiple cell types and plasma (16, 22, 23), were evaluated in parallel to directly compare the relative anti-inflammatory effects elicited by increased CYP-mediated EET biosynthesis and decreased sEH-mediated EET hydrolysis. Compared to wild-type controls, Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice exhibited a significant attenuation of endotoxin-induced activation of NF-κB signaling; CAM, chemokine and cytokine expression; and neutrophil trafficking in lung in vivo. Furthermore, attenuation of LPS-induced NF-κB activation and CAM and chemokine expression was observed in primary pulmonary endothelial cells isolated from Tie2-CYP2J2-Tr and Tie2-CYP2C8-Tr mice. Notably, this attenuation of the inflammatory response to LPS was inhibited by both an EET antagonist (14,15-EEZE) and a selective CYP epoxygenase inhibitor (MS-PPOH), directly implicating CYP epoxygenase-derived EETs to the observed anti-inflammatory phenotype. Collectively, these data demonstrate that potentiation of the CYP epoxygenase pathway by either increased endothelial EET biosynthesis or globally decreased EET hydrolysis attenuates NF-κB-dependent vascular inflammatory responses in vivo.

It is well established that NF-κB mediates transcriptional activation of cytokine, chemokine, and CAM expression (2–4). This is particularly important in the vasculature, where cytokines and chemokines attract inflammatory cells and CAMs facilitate adherence to endothelial cells and transmigration into tissue, such as the alveolar space in lung and intima in arteries, where they ultimately elicit local pathological damage. A series of in vitro studies have demonstrated that direct application of EETs (0.1–1 μM) or overexpression of CYP2J2 attenuates NF-κB activation and expression of downstream genes in endothelial cells (8, 15, 39–41) via inhibition of IκB kinase activity, IκBα phosphorylation, and degradation, and subsequent RelA nuclear translocation (15). We observed that LPS-induced IκBα phosphorylation and cytokine, chemokine, and CAM expression was similarly attenuated in Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice, compared to wild-type controls, demonstrating that potentiation of the CYP epoxygenase pathway, both systemically and locally in the endothelium, attenuates NF-κB activation in vivo. Wild-type C57BL/6, Tie2-CYP2J2-Tr and Tie2-CYP2C8-Tr, and Ephx2^−/− mice exhibit plasma 11,12-EET concentrations of ∼0.6 nM, 1 nM, and 2 nM, respectively. Although EETs are predominantly stored in cellular membranes via esterification to phospholipids, and intracellular concentrations may be higher than those observed in plasma (42), these concentrations are considerably lower than the supraphysiological concentrations utilized in the aforementioned in vitro experiments (15, 39, 40). Consequently, the anti-inflammatory effects observed in the current investigation occurred in the presence of physiological increases in EET levels, and demonstrate an important role for the CYP epoxygenase pathway in the regulation of inflammation in vivo.

Recent work has demonstrated the presence of Tie2 expression in distinct monocyte lineages and that the Tie2-driven transgenic expression is not exclusively limited to endothelial cells (43). Consequently, we cannot exclude the possibility that CYP2J2 and CYP2C8 expression in other cell types contributed to the observed anti-inflammatory phenotype in Tie2-CYP2J2-Tr and Tie2-CYP2C8-Tr mice in vivo. However, a similar attenuation of CAM and chemokine expression and IκBα phosphorylation was also observed in primary endothelial cells isolated from Tie2-CYP2J2-Tr and Tie2-CYP2C8-Tr mice, suggesting that endothelial CYP epoxygenase expression most likely drove the observed anti-inflammatory phenotype in vivo. Notably, the EET antagonist 14,15-EEZE blocked this effect, demonstrating the direct contribution of EETs to the attenuation of NF-κB activation elicited by endothelial CYP2J2 and CYP2C8 expression.

Acute lung injury is the leading cause of mortality in intensive care unit patients, and is an important consequence of septic shock (44). Consequently, new therapeutic strategies that ameliorate inflammation-induced lung injury could offer substantial clinical benefit. Prior studies have demonstrated that local and systemic exposure to endotoxin results in prominent activation of NF-κB signaling in the vasculature and airways, and NF-κB is an integral regulator of pulmonary inflammation and injury (5, 6). A recent study found that CYP-derived EETs and DHETs are the primary arachidonic acid metabolic products released on microbial challenge in isolated, perfused human lungs (45), suggesting that the CYP epoxygenase may serve an important functional role in the regulation of infection-mediated pulmonary inflammation and injury in humans. Moreover, pulmonary CYP epoxygenase expression and EET formation was significantly suppressed in rodent models of acute Pseudomonas pneumonia (46, 47) and endotoxin (48), further suggesting that dysregulation of CYP-mediated EET biosynthesis may contribute to the pathological response to inflammation in lung. Consequently, potentiation of CYP-derived EETs may be a rational therapeutic strategy to attenuate pulmonary inflammation. Indeed, previous studies have demonstrated that 14,15-EET administration attenuated cytokine-induced IκBα degradation in a primary culture of human bronchi (49), and administration of an sEH inhibitor attenuated tobacco smoke-induced inflammatory cell infiltration into bronchoalveolar lavage fluid in rats (50). The current investigation further demonstrates the potent anti-inflammatory effects of EETs, through both increased CYP-mediated biosynthesis and decreased sEH-mediated hydrolysis, in a model of endotoxemia-induced acute pulmonary inflammation.

Neutrophils are a major component of inflammatory cells in lung. MPO is predominantly localized in clusters of infiltrated and activated neutrophils and is an important functional mediator of neutrophil trafficking into lung and acute lung injury (37, 51). Consequently, quantification of MPO-expressing cells and enzymatic activity in the lung parenchyma serve as an in vivo marker of neutrophil infiltration (38). In the current study, induction of alveolar wall infiltration of MPO-positive cells and MPO functional activity by LPS were significantly attenuated in Tie2-CYP2J2-Tr, Tie2-CYP2C8-Tr, and Ephx2^−/− mice compared to wild-type controls. Collectively, these data suggest that potentiation of CYP-derived EETs inhibits neutrophil infiltration into lung via attenuation of NF-κB activation and CAM, chemokine, and cytokine expression. However, the overall extent of neutrophil infiltration in response to LPS was generally mild, representing a limitation of the current study. Although NF-κB signaling and proinflammatory gene expression are substantially up-regulated 3 h after LPS administration, induction of inflammatory cell infiltration initiates as early as 1–2 h and peaks ∼24 h after intraperitoneal LPS administration (20, 21, 52). Moreover, because of the transient nature of neutrophil infiltration following systemic LPS administration, relatively minor lung injury occurs in C57BL/6 mice even at near-lethal doses (20, 52). Notably, the current study rigorously characterized the molecular events that underlie LPS-induced neutrophil infiltration. However, additional studies will be necessary to fully characterize the effect of CYP-derived EETs on the kinetics of neutrophil trafficking and the extent of lung injury in vivo, and determine whether therapeutically increasing EETs represents an effective anti-inflammatory strategy to attenuate sepsis-related acute lung injury. Future studies in models of acute and chronic inflammation involving other organ systems also remain necessary to characterize the potential therapeutic utility of increasing CYP-derived EETs in other diseases in which inflammation plays an integral pathological role.

In summary, using three distinct mouse models, we demonstrated in the current study that increased endothelial CYP2J2 and CYP2C8-mediated EET biosynthesis and globally decreased sEH-mediated EET hydrolysis elicited a significant attenuation of endotoxin-induced activation of NF-κB signaling; CAM, chemokine, and cytokine expression; and neutrophil infiltration in lung. Collectively, our data demonstrate that potentiation of the CYP epoxygenase pathway attenuates acute, NF-κB-dependent vascular inflammatory responses in vivo, and may serve as a viable anti-inflammatory therapeutic strategy.