Role of Cytochrome P450 2C Epoxygenases in Hypoxia-Induced Cell Migration and Angiogenesis in Retinal Endothelial Cells

Abstract

purpose. Cytochrome P450 (CYP) epoxygenase-derived epoxyeicosatrienoic acids (EETs) elicit cell proliferation and promote angiogenesis. The aim of this study was to determine the expression of CYP epoxygenases in the bovine retina and the potential role of EETs in hypoxia-induced angiogenesis in bovine retinal endothelial cells.

methods. Bovine retinal endothelial cells were cultured under normoxic (21% O₂) or hypoxic (1% O₂) conditions, and CYP2C expression was determined by Western blot analysis. The effect of hypoxia on EET levels was determined by LC-MS/MS. Cell migration (Transwell filter assays) and endothelial cell tube formation (on basement membrane matrix) were assessed in vitro in the absence and presence of pharmacologic inhibitors and CYP2C antisense oligonucleotides.

results. Bovine retinal endothelial cells expressed CYP2C protein in culture and generated detectable levels of EETs under basal conditions. Hypoxia (6–48 hours) enhanced CYP2C protein expression (2-fold) and EET formation (1.5-fold). Moreover, endothelial cells preexposed to hypoxia demonstrated an increase in serum-induced cell migration that was sensitive to the CYP2C inhibitors sulfaphenazole and MS-PPOH and the EET antagonist 14,15-epoxyeicosa-5(Z)-enoic acid. Furthermore, preventing the hypoxia-induced expression of CYP2C (antisense oligonucleotides) suppressed hypoxia-induced cell migration. In an in vitro angiogenesis model, the preexposure of endothelial cells to hypoxia increased CYP2C expression and enhanced endothelial tube formation, which was blocked by the EET antagonist and by the CYP2C antisense oligonucleotides.

conclusions. Taken together, these data indicate that CYP2C-derived EETs are implicated in angiogenesis by retinal endothelial cells, especially under hypoxic conditions.

In the endothelium, cytochrome P450 (CYP) epoxygenases of the 2B, 2C, and 2J subfamilies have been reported to generate the four different regio-isomers (5,6-, 8,9-, 11,12-, and 14,15) of epoxyeicosatrienoic acid (EET) from arachidonic acid. Apart from their function as endothelium-derived hyperpolarizing factors, ^{1 2} some of the EETs exert additional membrane potential–independent effects on vascular cells and are now recognized as intracellular signaling molecules (for reviews, see Fleming ³ and Chen et al. ⁴ ). The increase in intracellular EET production results in effects ranging from changes in intracellular Ca²⁺ concentration ⁵ and in the activation of protein kinase A, ⁶ tyrosine kinases, and phosphatases ^{7 8} to the activation of mitogen-activated protein (MAP) kinase phosphatases and inhibition of the c-Jun N-terminal kinase. ⁹ Some of these responses may result from the direct interaction of an EET with the target molecule. For example, the binding of EETs to the TrpV4 channel may account for effects on Ca²⁺ signaling, ¹⁰ whereas others can be attributed to the transactivation of receptors such as the epidermal growth factor (EGF) receptor. ^{11 12} Several of the pathways activated by EETs have been demonstrated to lead to an increase in cell proliferation and angiogenesis (for review, see Michaelis and Fleming ¹³ ). However, because of the rapid decrease in CYP protein levels that occurs in cultured cells, most of the latter studies have been performed using overexpression systems or have monitored responses to the exogenous application of EETs. Consequently, little is known about the role of endogenously generated EETs in angiogenesis. To investigate the effects of endogenously expressed CYP epoxygenases, it is usually necessary to induce their expression. In addition to the well-known pharmacologic stimuli, ^{14 15 16} hemodynamic and physiochemical forces are able to enhance CYP2C expression in isolated endothelial cells. Indeed, we have previously reported that cyclic stretch induces CYP2C expression and activity as well as angiogenesis in human umbilical vein endothelial cells. ¹⁷ Another stimulus influencing the expression of several CYP enzymes is hypoxia, which downregulates the expression of CYP2J2, ¹⁸ whereas transient cerebral ischemia elicits the upregulation of CYP2C11 in rats ¹⁹ and the expression of CYP2C8/9 in human endothelial cells. ²⁰  

To date, the vascular effects of the EETs have been analyzed in a number of different organs, including the brain and the heart, where they protect against ischemia/reperfusion injury. ^{21 22 23} A role in the retina has been suggested on the basis of a recent report showing the involvement of EETs in neurovascular coupling, ²⁴ but nothing is known about CYP2C expression in the retina or, more specifically, the role of EETs in retinal angiogenesis. In the present study, we analyzed CYP2C expression in retinal endothelial cells and the role of CYP2C-derived EETs in hypoxia-induced cell migration and angiogenesis. 

Materials and Methods

Materials

Collagenase II was purchased from Worthington (Lakewood, NJ), and 14,15-epoxyeicosa-5(Z)-enoic acid (14,15-EEZE) and MS-PPOH were synthesized as described. ²⁵ Sulfaphenazole, the antibody recognizing β-actin, and all other chemicals were from Sigma-Aldrich (Steinheim, Germany). 

Immunofluorescence

Immunofluorescence analysis of vessel segments was performed as previously described. ²⁶ Proteins were detected using specific antibodies against CYP2C8 (Acris, Hiddenhausen, Germany), endothelial nitric oxide synthase (BD Transduction, Heidelberg, Germany), and α-actin (FITC-conjugated; Sigma-Aldrich). Fluorescent secondary antibodies (Alexa Fluor 546 goat anti-rabbit, Alexa Fluor 647 goat anti-mouse) were from Invitrogen Molecular Probes (Eugene, OR), and images were acquired by laser scanning microscopy (LSM 510 meta; Carl Zeiss, Jena, Germany). 

Cell Culture

Bovine retinal microvascular endothelial cells were isolated from calf eyes obtained from a local slaughterhouse, as described, ²⁷ and endothelial cell homogeneity was confirmed by the specific uptake of acetylated low-density lipoprotein. Endothelial cells were grown on dishes coated with collagen I and maintained in MEM containing 50 U/mL penicillin, 50 μg/mL streptomycin, endothelial cell growth supplement (30 μg/mL), and 10% horse serum from platelet-poor plasma. Endothelial cells between passages 4 and 10 were used in this study. Hypoxic conditions were achieved by incubating cells in an air-tight incubator (Heraeus, Hanau, Germany) with 5% CO₂ and 1% O₂ balanced with N₂. 

In some experiments, an antisense oligonucleotide approach was used to prevent the hypoxia-induced upregulation of CYP2C. Endothelial cells were treated with CYP 2C sense and antisense oligonucleotides (2 μM; antisense: 5′-TCC ATT GAA GCC TTC TCT TCT T-3′; sense: 5′-AAG AAGAGA AGG CTT CAA TGG A-3′; in both cases, the three 5′ nucleotides were modified with phosphothioate; MWG-Biotech, Ebersberg, Germany) using transfection reagent (GeneTrans II; MobiTec, Göttingen, Germany) according to the manufacturer’s protocol. ² The sequence of these oligonucleotides spans the ATG and is 100% identical with human CYP2C8 and contains one mismatch to the other three human CYP2C isoforms. As a consequence of the high homology between the 2C enzymes, ²⁸ the oligonucleotides used also targeted the bovine CYP2C isoforms. 

Immunoblotting

For Western blot analysis, cells were lysed in Triton X-100 lysis buffer and separated by SDS-PAGE, as described. ¹² The CYP2C antibody used was purified (Eurogenetec, Seraing, Belgium) from rabbits immunized with a CYP2C9 peptide (RRRKLPPGPTPLPIC). 

LC-MS/MS measurements

After exposure to hypoxia for 16 hours in the absence or presence of the pharmacologic agents described in Results, endothelial cells were cultured for another 2 hours under normoxic conditions to rule out any potential direct effect of low O₂ tension on CYP activity. Thereafter, the cells were harvested by scraping, and the cell pellets (from approximately 8 × 10⁶ cells) were suspended in 100 μL KCl (0.2 M) and 100 μL methanol, hydrolyzed for 2 hours using NaOH (0.5 N), neutralized with HCl (2 M), and spiked with deuterated internal standards (5-HETE-d₈, 12-HETE-d₈, 15-HETE-d₈, 20-HETE-d₆, 8,9-EET-d₈, 11,12-EET-d₈, and 14,15-EET-d₈). The liquid–liquid extraction was performed twice using 1 mL hexane/ethyl acetate (1:1). After evaporation of the solvent in a vacuum block under a gentle stream of nitrogen, samples were reconstituted with 50 μL methanol/water (1:1, vol/vol), and eicosanoids were determined with a mass spectrometer (API4000; MDS Sciex, Concord, ON, Canada) operating in multiple reaction-monitoring mode. Chromatographic separation was performed on an HPLC column (Gemini C18; 150 × 2 mm ID, 5-μm particle size, Phenomenex, Aschaffenburg, Germany). 

Cell Migration Assay

Endothelial cell migration was assessed using a modified Transwell chamber system with membrane inserts (FluoroBlok; 3-μm pore size; BD Biosciences, Heidelberg, Germany), as described. ²⁰ To stimulate cell migration, the lower chamber contained MEM supplemented with 10% horse serum and endothelial cell growth supplement, whereas the upper chamber contained MEM supplemented with 0.1% BSA. Control experiments revealed that no effect on cell adhesion or cell proliferation could be observed under these experimental conditions. 

In Vitro Angiogenesis Assays

To investigate the effect of hypoxia-induced CYP2C expression on tube formation, retinal endothelial cells were cultured in normoxic or hypoxic conditions for 24 hours in the absence and presence of 14,15-EEZE (10 μM). Thereafter, the cells were seeded onto basement membrane matrix (growth factor-reduced Matrigel; BD Biosciences) at a density of 1 × 10⁴ cells/cm² and cultured for another 12 hours under normoxic conditions, with and without 14,15-EEZE. Tube formation was assessed and quantified by counting the number of branch points. 

Statistical Analysis

Data are expressed as mean ± SEM. Statistical evaluation was performed with Student’s t-test for unpaired data, one-way ANOVA, and Bonferroni t-test or ANOVA for repeated measures where appropriate. P < 0.05 was considered statistically significant. 

Results

CYP2C Expression in the Bovine Retina

Immunohistochemical analysis of arterial segments freshly isolated from the bovine retina revealed the presence of CYP2C protein in the vascular endothelium. The CYP2C signal co-localized with that of the endothelial nitric oxide synthase, while no CYP2C protein was found in vascular smooth muscle cells that were labeled with α-actin (Fig. 1) . 

Effect of Hypoxia on CYP2C Expression and EET Formation in Bovine Retinal Endothelial Cells

Analysis of cultured bovine retinal endothelial cells revealed that CYP2C protein remains detectable after cell isolation; this contrasts with other cell types (porcine coronary artery endothelial cells, human umbilical vein endothelial cells) in which CYP2C protein and mRNA levels decrease rapidly in culture. ²⁹ Moreover, the exposure of cultured retinal endothelial cells to hypoxia significantly enhanced CYP2C expression (Fig. 2) . 

The relatively high basal expression of CYP2C protein correlated with the basal production of all four EET stereoisomers. The measured EET levels were markedly greater (approximately threefold) than those detected in primary cultures of human umbilical vein endothelial cells, which express very low CYP2C levels. Moreover, 11,12-EET and 14,15-EET levels were increased in retinal endothelial cells exposed to hypoxia for 16 hours (Fig. 3) . In contrast, there was no detectable effect of hypoxia on 20-HETE formation (data not shown). 

Effect of CYP2C Signaling Inhibition on Hypoxia-Induced Cell Migration

To investigate the involvement of CYP2C in hypoxia-induced endothelial cell migration, retinal endothelial cells were cultured under either normoxic or hypoxic conditions (24 hours) before they were seeded onto Transwell filters. Migration through the filter was examined after 20 hours of culture in the absence or presence of the CYP2C inhibitors MS-PPOH and sulfaphenazole or the EET antagonist 14,15-EEZE, respectively. Preexposure to hypoxia significantly enhanced cell migration, an effect that could be abolished by inhibiting EET production (using MS-PPOH, sulfaphenazole) or by antagonizing the actions of EET (14,15-EEZE; Fig. 4A ). To demonstrate a specific link to CYP2C proteins, we repeated the experiments using endothelial cells treated with CYP2C antisense oligonucleotides. Antisense oligonucleotides targeted against CYP2C only slightly attenuated basal CYP2C expression compared with cells treated with control (sense) oligonucleotides cultured under normoxic conditions (Fig. 4B) . However, antisense oligonucleotide treatment largely prevented the increase in CYP2C expression in response to hypoxia (1% O₂, 24 hours) and the subsequently assessed endothelial cell migration. 

Involvement of CYP2C in Hypoxia-Induced Tube Formation

Endothelial cell tube formation was assessed using the basement membrane matrix (Matrigel; BD Biosciences) assay. Retinal endothelial cells preexposed to hypoxia demonstrated augmented tube formation compared with cells cultured under normoxic conditions. The EET antagonist 14,15-EEZE slightly attenuated tube formation by normoxic cells but completely inhibited tube formation by cells cultured under hypoxic conditions (Fig. 5A) . Similarly, treatment of endothelial cells with CYP2C antisense oligonucleotides also prevented the hypoxia-induced formation of endothelial cell tubes (Fig. 5B) . 

Discussion

The results of the present investigation demonstrate that hypoxia induces CYP2C expression in bovine retinal endothelial cells and that hypoxia-induced angiogenesis is dependent on CYP2C expression/activity. Indeed, CYP2C antisense oligonucleotide treatment, CYP inhibition, and EET antagonism abolished hypoxia-induced endothelial cell migration and angiogenesis. 

Although endothelial cells are often assumed to display comparable features independent of their source, the results of the present study demonstrated a consistent expression of CYP2C protein in retinal endothelial cells (in situ and in vitro) that was largely unaffected by prolonged culture and passaging. This observation is in stark contrast to other endothelial cell types we have analyzed, in which CYP2C protein and mRNA levels decrease rapidly after cell isolation. ^{2 15} In accordance with CYP2C protein expression, EET levels in retinal endothelial cells are elevated compared with other cultured cells (human umbilical vein endothelial cells, porcine aortic endothelial cells). Thus, it appears that mechanisms specific to the retinal vasculature maintain the expression of CYP2C enzymes and that these are lacking or repressed in endothelial cells from other sources. Our observations cannot be attributed to the contamination of the retinal endothelial cell cultures by other CYP-expressing cells (e.g., astrocytes) because these were undetectable in the cultures used. 

The expression of several CYP enzymes can be modulated by changes in O₂ tension. On the one hand, hypoxia downregulates CYP2J2, ¹⁸ and, in patients with chronic pulmonary disease, the expression and activity of several drug-metabolizing CYP enzymes decrease after acute moderate hypoxemia. ³⁰ On the other hand, hypoxia induces the expression of the human epoxygenases CYP2C8 and CYP2C9, ²⁰ as well as their rat (CYP2C11) ^{19 31} and murine (CYP2C29 ³² homologues. The latter enzyme is thought to be so sensitive to hypoxia that it contributes to the phenomenon of acute hypoxia-induced vasoconstriction (Euler-Liljestrand reflex). ³² We have previously shown that hypoxia increases the activity of the CYP2C9 promoter and enhances the expression of CYP2C mRNA and protein in human and porcine endothelial cells. ²⁰ Given that increased CYP2C expression and activity in the latter cell types have been linked with increased endothelial cell proliferation and migration and with angiogenesis, a hypoxia-induced increase in CYP2C expression in bovine retinal endothelial cells can be justifiably linked to a similar process. Indeed, using a CYP2C antisense approach to prevent the increase in CYP expression essentially abrogated retinal endothelial cell angiogenesis. However, the induction of VEGF by hypoxia and the role of the latter growth factor in hypoxia-induced cell migration cannot be overlooked. Given the information currently available and our observation of almost complete reduction of hypoxia-induced cell migration after inhibition of the CYP2C signaling pathway, it is tempting to suggest a link between VEGF and EET signaling. Circumstantial evidence for such a link exists because VEGF has recently been reported to stimulate CYP activation in pulmonary arteries, ³³ and the angiogenic response initiated by the CYP4A product 20-HETE has been linked with an increase in VEGF expression. ³⁴  

Apart from generating EETs, some CYP enzymes (e.g., CYP2C9) are also able to generate 20-HETE and physiologically relevant amounts of superoxide anions (O₂ ⁻), ^{26 35} which are also important mediators of endothelial cell migration and angiogenesis. ^{4 34 36 37} Because the epoxygenase inhibitors MS-PPOH and sulfaphenazole and the CYP2C antisense treatment attenuate the production of EETs, 20-HETE, and CYP-derived O₂ ⁻, data obtained using these approaches alone cannot provide sufficient evidence to conclude a role for EETs in the regulation of hypoxia-induced retinal angiogenesis. Therefore, we assessed the effects of the EET antagonist, 14,15-EEZE, which prevents the effects of EETs without interfering with those of 20-HETE ²⁵ or the production of O₂ ⁻ by CYP2C enzymes. We found that 14,15-EEZE inhibited hypoxia-induced endothelial cell migration and tube formation, indicating that EETs and not reactive oxygen species are responsible for the effects observed. Moreover, we were unable to detect any increase in 20-HETE formation in endothelial cells exposed to hypoxia, indicating that 20-HETE does not contribute to angiogenesis under the conditions used. 

Not only is retinal angiogenesis a major process that occurs during development, it also underlies retinopathy of prematurity and advanced diabetic retinopathy, both of which cause irreversible blindness. The retina is a tissue particularly rich in polyunsaturated fatty acids, and it responds to physiological and pathophysiological stimuli by the activation of phospholipases and the consequent release from membrane phospholipids of biologically active metabolites. ³⁸ The diabetic state is characterized by the accumulation of advanced glycation end products that lead to the activation of cytosolic phospholipase A₂ and, therefore, the release of arachidonic acid. ³⁹ Because CYP2C activity is dependent on substrate availability, enhanced levels of arachidonic acid may lead to an augmented EET formation. Therefore, it is likely that diabetic conditions also increase retinal EET production. There is at least circumstantial evidence to support this hypothetical link because diabetes has been associated with changes in the expression of several CYP enzymes in different mouse models of the disease. In spontaneously diabetic TallyHo mice, CYP2C epoxygenase activity is reportedly elevated. Rather than being linked to vasodilatation, however, this effect was associated with endothelial dysfunction, an effect attributed to increased O₂ ⁻ formation and the generation of an endothelium-derived eicosanoid vasoconstrictor compound. ⁴⁰ Moreover, streptozotocin-induced diabetes in rats leads to differential regulation of CYP expression in the liver, with CYP1A2, 1B1, 2B1, and 2E1 upregulated and CYP2C11 protein almost completely abrogated. ⁴¹ In the same diabetes model, enhanced levels of vasodilator epoxides have been detected in the perfused kidney, where they are assumed to regulate vascular tone. ⁴² The situation in the retina is, however, likely to be more complicated because, in addition to endothelial cells, other cells such as astrocytes ^{43 44} and glial cells ²⁴ are also able to generate EETs, which in turn influence vascular function and homeostasis. Therefore, not only may EETs produced in endothelial cells contribute to retinal angiogenesis, those generated in glial cells and astrocytes may also play an important role. 

Taken together, the results of the present study demonstrate the expression of CYP2C epoxygenases in the retinal vasculature and the contribution of CYP-derived EETs to hypoxia-induced retinal angiogenesis. Given that hypoxia induces CYP2C expression and EET formation in the retina and subsequently EET-dependent angiogenesis, our results imply a potential role of EETs in diabetic retinal ischemia and retinopathy, a link that has to be more closely addressed in further studies. 

 Supported by the Deutsche Forschungsgemeinschaft Grant SFB-TR 23, A6, and by United States Public Health Service National Institutes of Health Grant GM31278.

 Submitted for publication August 20, 2007; revised October 23, 2007; accepted December 19, 2007.

 Disclosure: U.R. Michaelis, None; N. Xia, None; E. Barbosa-Sicard, None; J.R. Falck, None; I. Fleming, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

CYP2C expression in a retinal vessel. Representative image(s) showing the expression of CYP2C (red) in a retinal vessel. (A) Endothelial cells were labeled using antibodies directed against the endothelial nitric oxide synthase (blue) and vascular smooth muscle cells using α-smooth muscle actin (green). (B) No signal was observed in segments incubated with the secondary antibody alone. Identical results were obtained using material from two additional animals.

Figure 2.

 

Effect of hypoxia on CYP2C expression in bovine retinal endothelial cells. Western blot and bar graph comparing the effects of normoxia (N, 21% O₂) and hypoxia (H, 1% O₂) for up to 48 hours on CYP2C expression. Protein extract from bovine liver served as a positive control (pc). To demonstrate equal loading of each lane, membranes were reblotted with an anti–β-actin antibody. The bar graph summarizes data from four independent experiments. *P < 0.05 versus normoxia (N).

Figure 3.

 

Effect of hypoxia (1% O₂, 16 hours) and CYP epoxygenase inhibition on EET formation by bovine retinal endothelial cells. Intracellular EET content was measured by LC-MS/MS in endothelial cells exposed to normoxia or hypoxia over 16 hours and treated with either solvent (0.1% dimethyl sulfoxide) or MS-PPOH (MS, 10 μM) for 90 minutes. The bar graph summarizes the results obtained using four independent cell batches. *P < 0.05 versus normoxia in the presence of solvent (Sol).

Figure 4.

 

Effect of CYP2C signaling pathway inhibition on hypoxia-induced cell migration. (A) Effects of the CYP epoxygenase inhibitors MS-PPOH (MS, 10 μM) and sulfaphenazole (Sulfa, 10 μM) and the EET antagonist 14,15-epoxyeicosa-5(Z)enoic acid (EEZE, 10 μM) on the migration of bovine retinal endothelial cells preexposed to hypoxia (1% O₂, 24 hours) to increase CYP2C expression. (B) Effects of CYP2C downregulation on hypoxia-induced cell migration. Bovine retinal endothelial cells were transfected with control (CYP2C sense, Se) or CYP2C antisense (As) oligonucleotides and cultured under normoxic or hypoxic conditions for 24 hours. Thereafter, the cells were seeded onto Transwell filters, and migrated cells were counted after another 24 hours. Inset: representative Western blot demonstrating the effect of CYP2C antisense oligonucleotides and hypoxia on CYP2C expression. The bar graphs summarize the results of four independent experiments. *P < 0.05 versus normoxia in the presence of solvent (Sol) or sense oligonucleotides (Se).

Figure 5.

 

Effect of 14,15-EEZE and CYP2C antisense oligonucleotides on hypoxia-induced endothelial cell tube formation. (A) Bovine retinal endothelial cells were cultured under normoxic (21% O₂) or hypoxic (1% O₂) conditions in the absence or presence of 14,15-EEZE (EEZE, 10 μM) for 24 hours. Thereafter, the cells were seeded onto basement membrane matrix and cultured for 12 more hours before branching points were counted. (B) After transfection with CYP2C sense or antisense oligonucleotides, retinal endothelial cells were exposed to normoxia or hypoxia for 24 hours and seeded onto basement membrane matrix. The bar graphs summarize the results of three to four independent experiments. *P < 0.05 and **P < 0.01 versus normoxia in the presence of solvent (Sol) or sense oligonucleotides (Se).

The authors thank Alexandra Rueben for organizing the supply of bovine retinas and Isabel Winter for expert technical assistance.