Hypoxia Stimulates the Synthesis of Cytochrome P450-Derived Inflammatory Eicosanoids in Rabbit Corneal Epithelium1
- Christina Vafeas1,
- Paul A. Mieyal1,
- Ferdinando Urbano1,
- John R. Falck2,
- Kamlesh Chauhan2,
- Michael Berman1 and
- Michal Laniado Schwartzman1
- 1Department of Pharmacology, New York Medical College, Valhalla, New York (C.V., P.A.S., F.V., B.M., M.L.S.) and Departments of 2Biochemistry and Pharmacology, University of Texas Southwestern Medical Center, Dallas, Texas (J.R.F., K.C.)
Abstract
The corneal epithelium metabolizes arachidonic acid by a cytochrome P450-(CYP) mediated pathway to 12(R)hydroxy-5,8,10,14-eicosatrienoic acid [12(R)-HETE] and 12(R)hydroxy-5,8,14-eicosatrienoic acid [12(R)-HETrE]. Both metabolites possess potent inflammatory properties with 12(R)-HETrE being a powerful angiogenic factor and assume the role of inflammatory mediators in hypoxia- and chemical-induced injury in the cornea, in vivo. We developed an in vitro model of corneal organ culture to characterize the biochemical and molecular events involved in the increased synthesis of these metabolites. These cultured corneas exhibit epithelial cytochrome P450 CYP-dependent 12(R)-HETE and 12(R)-HETrE synthesis as indicated by chiral analysis and by the ability of CYP enzyme inhibitors to repress their synthesis. Hypoxia greatly and selectively stimulated the synthesis of 12(R)-HETE (7-fold over control normoxic conditions) and 12(R)-HETrE. The bacterial endotoxin, lipopolysaccharide, also increased the synthesis of these eicosanoids, substantiating the notion that this activity may function as an inflammatory pathway. These metabolites were detected in the culture medium by gas chromatography/mass spectroscopy (GC/MS) analysis and their levels significantly increased in hypoxia-treated corneas, further indicating their endogenous formation in response to injury. This in vitro model provides an excellent preparation for studying factors regulating the synthesis of these inflammatory eicosanoids and for isolating, identifying and characterizing the CYP protein responsible for their synthesis.
Acute injury to the cornea is marked by swelling and loss of transparency associated with conjuctival vasodilation, edema, neovascularization and the appearance of an inflammatory infiltrate in the cornea, conjuctiva and the tear film. Chronic inflammation is marked by an invasion of new blood vessels into the normally avascular cornea (Waring and Rodrigues, 1987). Epithelial eicosanoids have been implicated as inflammatory mediators responsible for initiation, development and progression of the inflammatory response (Williams and Higgs, 1988). Besides the generation of the well-known cyclooxygenase- and lipoxygenase-derived eicosanoids, the corneal epithelium metabolizes arachidonic acid via the CYP monooxygenase pathway to two potent inflammatory eicosanoids. 12(R)-HETE has been found to inhibit Na+/K+ATPase and to perturb hydration control in the cornea causing edema; 12(R)-HETrE, is a vasodilator, chemotactic and angiogenic factor. These metabolites were identified as CYP-derived eicosanoids based on the following: 1) their formation is localized to the microsomal fraction and is dependent on NADPH, an essential cofactor for CYP-mediated reactions; 2) their formation is not affected by aspirin, indomethacin, diclofenac, BW755C or NDGA (cyclooxygenase and lipoxygenase inhibitors) but is inhibited by inhibitors of CYP enzymes such as SKF-525A, clotrimazole and carbon monoxide and by antibodies to CYP or NADPH-CYP (c) reductase; 3) product stereoselectivity is different from lipoxygenase-derived hydroxeicosanoids, i.e., the R enantiomer, 12(R)-HETE is the predominant metabolite formed in the corneal epithelial preparations; 4) in vivo, the production of these metabolites by the corneal epithelium can be inhibited by depleting CYP proteins via the induction of heme oxygenase, a major regulatory enzyme for CYP activities (Laniado Schwartzman, 1997, and references therein).
That these metabolites, in particular 12(R)-HETrE, can assume the role of inflammatory mediators in the injured cornea is not only implied by their documented potent in vitro and in vivobiological activities but is also suggested from recent studies in which we have shown, using a closed eye-hydrogel contact lens model of anterior surface inflammation, that the corneal epithelial capacity to synthesize 12-HETE and 12-HETrE is greatly enhanced. This increase in metabolism was predominantly CYP-dependent and significantly correlated to the appearance and course of the inflammatory response, thereby, indicating a potential cause-effect relationship between these events (Conners et al., 1995). This relationship was also observed in a model of alkali burn injury where CYP-derived 12(R)-HETE and 12(R)-HETrE production positively correlated to the degree of inflammation and neovascularization (Conners et al., 1997).
The results of the previous studies in living animals are particularly meaningful in that the biological activities of 12(R)-HETE and 12(R)-HETrE reproduce many of the hypoxia-induced pathophysiological events seen with extended contact lens wear in humans, e.g., vasodilation, edema and neovascularization. However, studies of the effects of hypoxia using soft contact lenses are unwieldy and are sometimes complicated by corneal injury, infection and the need to suture the lids shut to avoid loss of the lenses. Moreover, thein vivo contact lens-induced hypoxia model requires a week or more of closed eye wear for the development of large amounts of pathophysiologic changes (Conners et al., 1995). Similarly, the alkali burn-induced injury also required approximately 7 days for full expression of similar biochemical and pathophysiological changes (Conners et al., 1997). Furthermore, tissue availability for biochemical and molecular analysis is a problem with these in vivo models. The current in vitro model was developed to avoid the problems associated with the in vivo models. This model of corneal organ culture allows the measurement of biochemical changes in response to different amounts of stimuli in a reproducible manner. In this study, we demonstrate that cultured corneas exhibit epithelial CYP-dependent 12(R)-HETE and 12(R)-HETrE synthesis and that this synthesis is markedly enhanced under hypoxic conditions and in response to inflammatory stimuli such as the bacterial toxin, lipopolysaccharide.
Materials and Methods
Materials
Rabbit eyes from 8- to 12-wk-old male and female New Zealand White/California White rabbits (1.6–2.5 kg) were obtained from Pel-Freez Biologicals (Rogers, AK). Fresh eyes were shipped overnight on wet ice in Hanks Balanced Salt Solution containing 100 μg/ml penicillin G potassium, 100 U/ml streptomycin sulfate and 0.25 μg/ml amphotericin B (Fungizone) (1× pen/strep/amphotericin B). DMEM (high glucose), FBS and antibiotics/antimycotic (penicillin/streptomycin/amphotericin B) were from Gibco/BRL (Grand Island, NY). LH, PMSF, clotrimazole, indomethacin, LPS, IL-I and indomethacin were obtained from Sigma Chemical Co. (St. Louis, MO). β-Naphthoflavone was from Aldrich Chemical Co. (Milwaukee, WI).d-Glucose-6-Phosphate:NAD 1 oxidoreductase, NADP and G6P were from Boehringer-Mannheim (Indianapolis, IN). 1-14C-arachidonic acid (55 mCi/mmol) was from NEN Du Pont (Boston, MA). [5,6,8,9,11,12,14,15-3H]-12(S)-HETE (103 Ci/mmol), [2H8]-12-HETE and [14,15-H3]-12(R)-HETrE (31.25 Ci/mmol) were from Amersham (Arlington Heights, IL). 12(R)-HETE, 12(S)-HETE, 17-octadecaenoyic acid and cinnamyl-3,4-dihydroxy-α-cyanocinnamate were from Cayman Chemical Co. (Ann Arbor, MI). 12(R)-HETrE, 12(S)-HETrE and2H3-12(R)-HETrE were chemically synthesized by Dr. J. R. Falck (University of Texas Southwestern Medical Center at Dallas) as previously described (Shin et al., 1989).
Cornea Organ Culture
Eyes received in HBSS and antibiotics were removed to a large volume of DMEM with 2× pen/strep/Amphotericin B for 10 min at room temperature with occasional stirring (250 ml/30 eyes), after which eyes were soaked in 200 ml DMEM with 1× pen/strep/amphotericin B. A scleral incision was made 2 to 3 mm beyond the limbus and the cornea was removed with the scleral rim. The cornea/scleral rim tissue was then transferred to a plastic Petri plate containing 10 ml of DMEM with 1× pen/strep/amphotericin B (15 corneas per plate). The corneas were washed five times with the same medium and each cornea was transferred to a well of a 12-well plate containing 1 ml of DMEM and 1× pen/strep/amphotericin B with either 17% FBS or 0.2% LH; hence, the corneal epithelial surface was wetted by a thin film of medium, with the epithelium at the gas-liquid interface. Culture plates were then placed in a 37°C incubator supplied with 5% CO2/95% air (∼20% O2) (Normoxia), or in a modular tissue culture chamber (Billups-Rothenburg; DelMar, CA) supplied continuously with 5% CO2/2% O2/93% N2 (hypoxia) and bubbled through deionized H2O into the chamber within a 37°C incubator. Cultures went for 24 to 48 hr without change of medium.
Arachidonic Acid Metabolism
At the end of the culture period, culture plates were placed quickly on wet ice; the epithelium from each cornea was scraped into 500 μl of prechilled potassium phosphate buffer, pH 7.4 containing 0.1 mM PMSF in a prechilled Petri dish. Pooled epithelial scrapings (four corneas) were then collected into a 1.5 ml Eppendorf tube that was concentrated by centrifugation for 5 min at 3500 ×g, at 4°C. Supernatants were removed and pellets were resuspended in 500 μl cold potassium phosphate buffer and homogenized using a glass homogenizer. Ten μl were removed from each resultant homogenate for subsequent protein assay (Bradford, 1976) and known volumes of homogenate (320 μl, 160–300 μg protein) were then added to reaction mixtures containing the following components in a final volume of 500 μl of KP/PMSF buffer, pH 7.4: ±2.9 U G6P-DH, 3.3 mM G6-P, 0.5 mM NADP, 5 mM MgCl2. Reactions were started by the addition of [1-14C]-arachidonic acid (0.8 μCi, 28 μM). Incubation was for 1 hr at 37°C with constant shaking. In some experiments, inhibitors were added at the indicated concentration to the reaction mixture 15 min before the addition of arachidonic acid. Reactions were stopped by acidification with 0.2 M formic acid to pH 4.0 and metabolites were extracted with ethyl acetate as previously described (Conners et al., 1997). The final extract was dissolved in methanol and subjected to RP-HPLC separation of14C-arachidonic acid metabolites.
Separation of Arachidonic Acid Metabolites by RP-HPLC
Separation of the metabolites was performed by RP-HPLC in a Hewlett Packard 1050 system on a 5 μm ODS-Hypersil column, 4.6 × 200 mm (Hewlett Packard, Palo Alto, CA) using a solvent composed of 80% acetonitrile/water/acetic acid, 50:50:0.1 (v/v/v) and 20% acetronitrile/acetic acid, 100:0.1 (v/v), at a flow rate of 1 ml/min for 25 min followed by 100% acetonitrile/acetic acid, 100:0.1 (v/v) for 10 min. Radioactivity was monitored by an on-line flow detector (Radiometric Instruments, Tampa, FL). Identification of metabolites was based on comigration with authentic standards and GC/MS analysis as previously described (Conners et al., 1996).
Chiral analysis
For purposes of resolution and determination of the relative amounts of the S and R enantiomers of 12-HETE and 12-HETrE, separate fractions containing 14C-labeled 12-HETE or 12-HETrE were collected from RP-HPLC effluents. The samples, synthetic 12-HETE and 12-HETrE enantiomers, and [3H]-12(S)-HETE and [3H]-12(R)-HETrE were methylated using ice-cold, freshly prepared diazomethane for 5 min at room temperature. The methyl ester derivatives were separated on a Chiracel OC column (Daicel, Exton, PA) using hexane:isopropanol (98:2 v/v) at 0.5 ml/min. Radioactivity was detected by an on-line flow radioactive detector as previously described (Conners et al., 1996).
NCI-GC/MS
Measurements of 12-HETE.
3H-12-HETE (50,000 cpm) for extraction efficiency and [2H8]-12-HETE (10 ng) as an internal standard were added to the corneal organ culture medium after 24 hr under normoxic or hypoxic conditions. The medium was acidified and metabolites extracted with ethyl acetate as described above. The extract was subjected to RP-HPLC separation and fractions containing 12-HETE and 12-HETrE were collected. 12-HETE was derivatized to the PFB TMS as described (Laniado Schwartzman et al., 1991). NCI-GC/MS was performed on a Hewlett Packard HP 5989A mass spectrometer interfaced with a SP-2330 capillary column programmed from 100 to 250°C at 25°C/min using helium as the carrier gas. The mass spectrometer was scanned from 200 to 600 mass units in one second or single ions were monitored, m/z 391 for the endogenous derivatized 12-HETE (PFB ester TMS ether) and m/z 399 for the derivatized internal standard. Determination of the total 12-HETE content in the purified biological sample was made by comparison of the ion intensities at m/z 391:399 vs. a standard curve of 12-HETE-PFB-TMS/2H8-12-HETE-PFB-TMS molar ratio constructed by NCI-GC/MS analysis.
Measurements of 12-HETrE.
The procedure is essentially the same as that described for 12-HETE. The radioactive tracer is 14,15-3H-12(R)-HETrE (31 Ci/mmol) and the internal standard is 2H3-12(R)-HETrE. The mass spectrometer was scanned from 200 to 600 mass units in one second or single ions were monitored, m/z 393 for the endogenous derivatized 12-HETrE (PFB ester TMS ether) and m/z 396 for the derivatized internal standard. Determination of the total 12-HETrE content in the purified sample was made by comparison of the ion intensities atm/z 393:396 vs. a standard curve of 12-HETrE-PFB-TMS/2H3-12-HETrE-PFB-TMS molar ratio constructed by NCI-GC/MS analysis.
Data and Statistical Analysis
Specific activity was calculated from the percent conversion of arachidonic acid and expressed as nmol/mg protein/hr. Tests of significance of differences between specific activities were made using the modified Wilcoxon nonparametric Kruskal-Wallis one-way analysis of variance.
Results
Characterization of arachidonic acid metabolism in epithelium from cultured rabbit corneas.
The need for an in vitro model that could mimic the in vivo model of contact lens-induced hypoxic injury as well as other inflammatory insults provided the impetus for this current model. In this model, the cornea is incubated in culture medium containing either serum or LH in a humidified 37°C incubator. The corneal epithelium is then scraped and analyzed for its ability to metabolize arachidonic acid. Figure1 depicts a representative HPLC tracing of radioactive metabolites extracted from an incubate of epithelial homogenate with radioactive arachidonic acid and NADPH. Two major metabolites with retention times identical to those recovered from rabbit corneal microsomes (Conners et al., 1995) or freshly obtained corneal epithelium from either contact lens or alkali burn treated eyes (Conners et al., 1995, 1997) were detected. These two metabolites were identified as 12-HETE and 12-HETrE by their comigration with authentic standards. The production rates of 12-HETE and 12-HETrE by the epithelium after 24 hr in a culture medium devoid of serum was about two to three times higher than that of microsomal preparation from freshly isolated corneal epithelium (1.78 ± 0.35 nmol 12-HETE + 12-HETrE/mg/hr) (Stoltz et al., 1994), suggesting that the 24 hr in culture produced an increase in the baseline production. The activity was not significantly different when corneas were cultured with FBS in place of LH (LH, 9.66 ± 1.72 and 1.66 ± 0.24 nmol/mg/hr 12-HETE and 12-HETrE, respectively; FBS, 7.91 ± 0.33 and 0.96 ± 0.08 nmol/mg/hr 12-HETE and 12-HETrE, respectively). The activity under these conditions was maximal at 24 hr and decreased thereafter (vide infra). Therefore, experiments were generally carried out after a period of 24 hr in culture medium containing LH.
A representative reverse-phase HPLC separation of radioactive arachidonic acid metabolites formed by the epithelium from corneas in culture under normoxic conditions. Corneas were placed in 12-well culture plates in 1 ml DMEM containing 0.2% LH for 24 hr at 37°C incubator supplied with 5% CO2/95% air. The corneal epithelium was scraped, homogenized, incubated with14C-arachidonic acid and NADPH for 1 hr at 37°C, and metabolites extracted and separated by HPLC as described in “Materials and Methods.”
Effect of hypoxia on 12-HETE and 12-HETrE synthesis.
Hypoxia is thought to be the primary stimulus for the increased production of the CYP-derived 12-hydroxyeicosanoids in the corneal epithelium after contact lens wear, although mechanical and/or inflammatory stimulation cannot be ruled out. Thus, we used hypoxia as the stimulus in thein vitro organ culture which avoided mechanical and/or inflammatory stimuli. Incubation under hypoxic conditions (5% CO2/2% O2/93% N2) for 24 hr markedly enhanced the production of these two eicosanoids to a level two to three times more than the control, normoxic condition, 5% CO2/95% air (∼20% O2) (fig.2). Incubation for 48 hr under normoxic or hypoxic conditions resulted in a marked decrease in arachidonic acid conversion to 12-HETE (65% decrease) (fig. 2). Although the time-dependent decrease in 12-HETE formation under hypoxia was similar in magnitude to that observed under normoxia, significant 12-HETE levels were still produced, suggesting that inducing conditions such as hypoxia can maintain this activity in culture for a longer period of time. There was no significant time-dependent effect on 12-HETrE production (fig. 2). Furthermore, when corneas were cultured for 24 hr under normoxic conditions and then transferred to the hypoxic conditions for additional 24 hr, a significant 50% increase in 12-HETE synthesis was observed (14.97 ± 0.59 nmol/mg/hr as compared to 10.55 ± 0.39 nmol/mg/hr after 48 hr in normoxic conditions).
Effect of hypoxia on corneal epithelial synthesis of 12-HETE and 12-HETrE. Corneas were placed in 12-well culture plates in 1 ml DMEM containing 0.2% LH for 24 to 48 hr at 37°C incubator supplied with 5% CO2/95% air or 5% CO2/2% O2/93% N2. The corneal epithelium was scraped, homogenized, incubated with 14C-arachidonic acid and NADPH for 1 hr at 37°C, and metabolites extracted and separated by HPLC as described in “Materials and Methods.” Results are mean ± S.E., n = 8 for each treatment. * P < .001 from time-matched normoxia; ‡ P < .008 from 24 hr; P < .001 from 48 hr normoxia and 24 hr hypoxia.
That these metabolites, at least in part, are CYP-derived is indicated by two sets of experiments. First, addition of 17-ODYA or clotrimazole, inhibitors of CYP-derived arachidonic acid metabolism (Ortiz de Montellano and Reich, 1984; Capdevila et al., 1988), to the incubation mixture inhibited arachidonic acid conversion to 12-HETE by 41 and 35%, respectively, and to 12-HETrE by 37 and 28%, respectively (table 1). Second, chiral analysis indicated the presence of the R enantiomer for each of these metabolites (fig. 3). Under control, normoxic conditions, the R enantiomer constitutes 15% of total 12-HETE formed. However, under hypoxic conditions 26% of total 12-HETE formed was the R enantiomer, indicating a selective 7-fold increase in the formation of 12(R)-HETE (1.15 ± 0.3 and 7.57 ± 2.45 nmol/mg/hr in normoxia and hypoxia, respectively, mean ± S.E.,n = 7). Although the majority of 12-HETE was the S enantiomer, its fold increase in response to hypoxia was less than that of the R enantiomer (6.64 ± 0.7 and 21.20 ± 1.2 nmol/mg/hr, for normoxia and hypoxia, respectively, mean ± S.E.,n = 7) (fig. 3). As for the enantioselectivity of 12-HETrE, we found that under conditions of normoxia or hypoxia 85 ± 5% was in the R form and 16 ± 5% (mean ± S.E.,n = 3) was present as the S enantiomer. Lipoxygenase activity is highly stereoselective, resulting in the formation of 12(S)-HETE; cytochrome P450 monooxygenases are less stereospecific in that both 12(S)-HETE and 12(R)-HETE can be formed (Capdevila et al., 1986). To that end, CYP, a potent lipoxygenase inhibitor (IC50 = 0.063 for 12-lipoxygenase (Cho et al., 1991)), at 0.5 μM inhibited hypoxia-induced 12-HETE formation by 38% (fig. 3A) suggesting the presence of 12-lipoxygenase activity as well. CDC had no effect on 12-HETrE formation (table 1). The cyclooxygenase inhibitor, indomethacin (IC50 = 5 μM), did not affect 12-HETE formation. However, indomethacin had a significant effect on 12-HETrE formation reducing its synthesis by 45% (table 1) and suggesting that the oxidation-keto-reduction steps involved in the formation of 12-HETrE may be indomethacin-sensitive.
Effect of inhibitors of arachidonic acid metabolism on hypoxia-induced corneal epithelial synthesis of 12-HETE and 12-HETrE
Effect of hypoxia on the synthesis of 12-HETE enantiomers. A, A representative chiral HPLC chromatogram of 12-HETE enantiomers formed in the corneal epithelium from corneas incubated under hypoxic conditions. B, Chiral separation of 12(R)-HETE and 12(S)-HETE standards. C, Bar graph showing the effect of hypoxia on corneal epithelial 12(R)-HETE and 12(S)-HETE synthesis. Incubations and chiral analysis were carried out as described in “Materials and Methods.”
Effect of LPS on corneal epithelial synthesis of 12-HETE and 12-HETrE.
The bacterial endotoxin, LPS, is considered a general inflammatory stimulus to many cells. The CYP-derived metabolites 12(R)-HETE and 12(R)-HETrE not only exhibit inflammatory properties in vitro, but also have been shown to be produced, in vivo, in response to injury and during the inflammatory response. We examined whether an inflammatory stimulus alone can activate the synthesis of these powerful inflammatory mediators in corneal epithelium by administering LPS to the culture medium. Addition of LPS to the incubation medium of the cultured corneas resulted in a significant increase in the synthesis of these metabolites (fig.4). The synthesis of 12-HETrE increased by 3-fold over the control although that of 12-HETE increased by 2-fold. We also tested the effect of IL-1, one of the cytokines which has been shown to increase in many cells types after exposure to LPS and is considered to be an inflammatory cytokine. IL-1 did not significantly stimulate 12-HETE or 12-HETrE production over the control levels (fig. 4) suggesting that the mechanism by which LPS stimulates the synthesis of these eicosanoids is independent of IL-1.
Effect of LPS and IL-1 on corneal epithelial synthesis of 12-HETE and 12-HETrE. Corneas were placed in 12-well culture plates in 1 ml DMEM containing 0.2% LH and LPS (10 μg/ml) or IL-1 (20 ng/ml) for 24 hr at 37°C incubator supplied with 5% CO2/95% air or 5% CO2/2% O2/93% N2. The corneal epithelium was scraped and homogenized. Homogenates were incubated with 14C-arachidonic acid and NADPH for 1 hr at 37°C, and metabolites were extracted and separated by HPLC as described in “Materials and Methods.”
Detection of 12-HETE and 12-HETrE in corneal incubation medium.
Our results demonstrate the ability of the corneal epithelium to metabolize exogenously added arachidonic acid to 12(R)-HETE and 12(R)-HETrE under conditions of appropriate stimulation. However, they do not indicate endogenous production of these metabolites. We developed an NCI/GC/MS assay to determine whether these metabolites could be detected in the incubation medium of these corneas. The results indicated that both 12-HETE and 12-HETrE are released into the medium in ng amounts and that their levels are increased many folds after 24 hr in hypoxic conditions (fig.5). 12-HETE levels increased from 0.22 ± 0.06 to 4.28 ± 0.99 ng/ml and 12-HETrE from 0.012 ± 0.005 to 0.36 ± 0.08 ng/ml. Although chirality could not be determined by this method, it did demonstrate the capacity of the corneas to generate these metabolites from endogenous sources of arachidonic acid.
Effect of hypoxia on 12-HETE and 12-HETrE levels in the medium of corneal organ cultures. Corneas were placed in 12-well culture plates in 1 ml DMEM containing 0.2% LH for 24 hr at 37°C incubator supplied with 5% CO2/95% air or 5% CO2/2% O2/93% N2. At the end of the incubation period, the medium was pooled from four corneas and 12-HETE and 12-HETrE internal standards were added. Metabolites were extracted and separated by RP-HPLC as described in “Materials and Methods.” HPLC fractions containing 12-HETE and 12-HETrE were derivatized and further subjected to SIM-CG/MS analysis as described. Results are from three separate experiments. The levels of 12-HETE and 12-HETrE in hypoxic culture medium were significantly different from normoxic cultures, P < .01.
Discussion
Numerous studies have described the potent inflammatory properties of 12(R)-HETE and 12(R)-HETrE, implicating them as inflammatory mediators in hypoxic and chemical injury of the cornea (Laniado Schwartzman, 1997, and references therein). Edema and corneal neovascularization are potentially serious clinical problems associated with extended contact lens wear in humans (Holden, 1989) and are generally attributed to contact lens-induced hypoxia. However, regarding such edema and angiogenesis, a contribution derived from mechanical trauma inflicted by the contact lenses, in vivo, to the etiology of the pathological changes cannot be ruled out. Our work was undertaken to develop and characterize an in vitroorgan culture system for continuing studies of the possible role(s) of the CYP pathway in the rabbit corneal inflammatory response to injury such as hypoxia. The advantages of such a system are that environmental conditions can be controlled (e.g., hypoxia can be induced under defined O2 tensions and without mechanical injury or infections as contributing factors); and other potential modulators of CYP can be screened rapidly, without the problems associated with drug delivery in vivo.
In our study, hypoxia was produced by supplying the cultured corneas with 5% CO2/2% O2/93% N2; hypoxia was defined as 2% O2 because, clinically, O2 tensions less than 10% have been reported to cause edema, and O2 tension under hard contact lenses is ∼0 (Holden, 1989). Hypoxia resulted in 3- and 2-fold increases in the specific activity of 12-HETE and 12-HETrE formation, respectively. The observation that 12-HETE and 12-HETrE are produced in vitro, under hypoxic conditions in the absence of contact lenses suggests that contact lens-related trauma contributed to a lesser degree than hypoxia to the increases in 12-HETE and 12-HETrE reported in a previously published study on the effects of extended soft contact lens wear in rabbits (Conners et al., 1995). Noteworthy is the fact that the 12-HETE and 12-HETrE production by the epithelium of these cultured corneas under normoxic conditions was markedly higher than that obtained from freshly prepared epithelial homogenates (0.07 ± 0.02 nmol/mg/hr) or microsomes prepared from untreated Pel-Freeze eyes (1.78 ± 0.35 nmol/mg/hr) that, unlike the eyes for organ cultures, are shipped on ice (Conners et al., 1995, 1997). This suggest that the culture conditions alone have an inducing effect on the capacity of the epithelium to form these metabolites. One reason may lie in the fact that the corneas in 1 ml of medium was sometimes folded back on themselves so that the cultures had a variable geometry with regard to the gas phase. Because it has been reported that the distance from the gas phase is a critical determinant of the availability of O2 for mammalian tissue in culture fluid (Moscona et al., 1965), the variable geometry observed was a concern. We have modified the system to simulate the in vivosituation by placing the corneas on top of plastic domes eliminating possible variation in geometry. In this arrangement, 1 ml of culture medium just covered the corneal epithelium and, thus, provided a very thin aqueous layer between the epithelium and the gas phase. Cultured in this way, corneas showed an increase in epithelial formation of 12-HETE and 12-HETrE after hypoxia. However, the activity was one-half that measured in epithelium from submerged corneal cultures but similar to that obtained from freshly isolated epithelium (data not shown), suggesting a direct relationship between O2 concentration and CYP-derived eicosanoid synthesis.
As a further modification of the system, cultures were done in which LH, as a peptide source, was substituted for FBS; because, possibly, serum components, e.g., growth factors, variable from batch-to-batch of serum, could affect levels of CYP in the epithelium. It has been shown that cytokines, including growth factors, inhibit cytochrome P450 enzyme activities in several tissues (Chen et al., 1992, 1995). However, serum is frequently added to cultured cells for the maintenance of normal growth and function in vitro. Our results, however, indicated that corneas cultured in DMEM containing LH in the absence of serum exhibited similar, slightly higher, levels of epithelial 12-HETE and 12-HETrE synthesis as compared to corneas cultured in the presence of serum. Another parameter that may influence CYP activity in culture is time. Many cells demonstrate a time-dependent rapid loss of CYP enzyme activity and expression after cultivation (Paine, 1991) which can be maintained, to some extent, with CYP inducers or cofactors (Thurman and Kauffman, 1980). Indeed, we observed a marked loss of 12-HETE formation in epithelium from normoxia-cultured corneas within 48 hr; however, under hypoxia, 12-HETE production, although significantly decreased with time, remained higher compared to that in normoxia. These results suggest that the experimental inducer, i.e., hypoxia, as with other CYP inducers, is capable of maintaining significant enzymatic activity at least up to 48 hr in culture. Interestingly, the formation of 12-HETrE did not change with time. The formation of 12-HETrE requires the activity of additional enzymes downstream of CYP monooxygenase; thus 12-HETrE can be formed from CYP-derived 12(R)-HETE or 12(S)-HETE as well as from lipoxygenase-derived 12(S)-HETE via an oxidation-keto-reduction pathway (Yamamoto et al., 1994). These activities may compensate for the loss of CYP with incubation time.
That CYP activity is present and is involved in the production of epithelial 12-HETE and 12-HETrE in the cultured corneas is suggested by the inhibitory effect of the CYP enzyme inhibitors, clotrimazole and 17-ODYA and the presence the CYP-derived R enantiomer, 12(R)-HETE (Capdevila et al., 1986; Schwartzman et al., 1987). Moreover, hypoxia resulted in a greater increase in the R enantiomer, further implicating the involvement of a hypoxia-induced CYP activity. Others (Asakura et al., 1994; Lin et al., 1993) have reported that pig corneal epithelial microsomes do synthesize NADPH-dependent 12(R)-HETE, as well as 12(S)-HETE, from arachidonic acid, indicating the presence of both CYP monooxygenase and 12-lipoxygenase activities. Hurst et al. (1991) demonstrated the activation of 12-lipoxygenase in the cryogenically and alkali-burn injured rabbit corneas leading to increased levels of 12(S)-HETEin vitro. The inhibition of 12-HETE production by CDC may reflect the contribution of 12-lipoxygenase to the overall activity of 12-HETE synthesis in rabbit corneal epithelium after hypoxic injury. Whether CYP-derived or lipoxygenase-derived, the two enantiomers, 12(S) and 12(R)-HETE, may end up in a common pathway within the corneal epithelium, via an oxidation-ketoreduction-generated intermediate that gives rise to the proinflammatory mediator 12(R)-HETrE (Nishimuraet al., 1991; Yamamoto et al., 1994). It is noteworthy that the production of 12(R)-HETrE is markedly increased under hypoxic conditions. Because it is a potent angiogenic factor and angiogenesis is frequently associated with hypoxic conditions, 12(R)-HETrE may mediate, in part, the hypoxia-induced neovascularization of the cornea, e.g., after contact lens wear (Conners et al., 1995).
The demonstration that these metabolites can be detected in the culture medium and that their levels increased in hypoxia-treated corneas, further demonstrated their endogenous formation in response to hypoxia. Moreover, the estimated concentrations of these metabolites were in the range of their biological activities (∼10 and 1 nM for 12-HETE and 12-HETrE, respectively). Experiments using rabbit aqueous humor after alkali-burn induced inflammation and human tears, clearly indicated that samples taken from inflammatory conditions (alkali burn-treated eyes and overnight human tears) have higher levels (2- to 4-fold) of these inflammatory molecules (Husted et al., 1997). The CYP-arachidonic acid system has been demonstrated to reside in the corneal epithelium, but it has been difficult to provide strong support for the endogenous existence and functional/physiologic significance of its various metabolites, including 12-HETrE. This study lends support to the existence of 12-HETrE in significant quantities in the cornea and, possibly, other tissues in the anterior chamber.
The formation of 12(R)-HETE and 12(R)-HETrE in the corneal epithelium constitutes an inflammatory pathway; it is induced shortly after an injury to the corneal epithelium and the levels of both metabolites correlate with the severity of the inflammatory response (Connerset al., 1995, 1997). Moreover, both metabolites possess potent biological activities characteristic of inflammatory mediators. Whether this activity represents a common pathway in the response of the corneal epithelium to injury is yet to be examined. However, our finding that LPS, a bacterial toxin and potent inflammatory stimulus, markedly induced the formation of epithelial 12-HETE and 12-HETrE, suggests that this pathway may function in mediating the corneal inflammatory response. This finding is intriguing since LPS, when injected to rats, has been shown to inhibit hepatic CYP-derived drug metabolism (Morgan, 1993). This inhibition has recently been attributed to the induction of nitric oxide synthase by LPS and the generation of large amounts of nitric oxide, an inhibitor of CYP via its binding to the heme prosthetic group (Carlson and Billings, 1996; Khatsenkoet al., 1993). However, a recent report by Sewer et al. (1997) demonstrated that not all CYP isoforms respond to LPS in the same manner. In fact, the expression (mRNA and activity) of the renal CYP4A gene family is increased following LPS administration (Sewer et al., 1997). Recent studies in our laboratory indicate the presence of a hypoxia-induced CYP isoform of the CYP4 gene family in the cornea (Mastyugin et al., 1998). Whether it is also induced by LPS is yet to be determined.
The formation of these metabolites is also seen in other tissues and their topical application elicits biological activities typical of inflammatory factors. Previous studies demonstrated the production of 12(R)-HETE, and possibly 12-HETrE, in human psoriatic lesions (Woollard, 1986; Camp et al., 1988; Camp et al., 1993). Human and rat epidermis have been shown to produce 12-HETE enantiomers by both lipoxygenase and CYP monooxygenase pathways, and rat epidermal microsomes have been shown to produce 12-HETrE enantiomers by an NADPH-dependent mechanism (Holtzman et al., 1989; Van Wauwe et al., 1991). Production of 12-HETrE from 12-HETE was also documented in porcine neutrophils (Wainwright et al., 1990). 12(R)-HETE is a chemoattractant for human neutrophils and mixed peripheral blood lymphocytes at high nM concentrations (Cunningham and Woollard, 1987; Bacon et al., 1988). It also causes erythema after topical administration of nmol doses to human skin (Woollard et al., 1988). In guinea pig skin, 12(R)-HETrE enhances delayed-type hypersensitivity inflammatory reactions, with a maximal 50% enhancement at 1 fmol doses (Connerset al., 1994). Proctor et al. (1989) have shown the formation of CYP-derived 12-HETE in the hamster cheek pouch after thermal injury and further demonstrated that 12(R)-HETE is a potent vasodilator in this inflamed tissue. The implication is clear: this pathway and its products may be involved in the inflammatory responses of tissues besides the cornea. Skin and corneal epithelium are derived from surface ectoderm and are well known to be clinically related. It is possible that these two tissues share a similar response to injury as well as a similar mechanism for regulating the inflammatory response.
Footnotes
-
Send reprint requests to: Dr. Michal Laniado Schwartzman, Department of Pharmacology, New York Medical College, Valhalla, NY 10595.
-
↵1 This work was supported by National Institutes of Health Grants EY06513 (M.L.S.) and DK38226 (J.R.F.).
- Abbreviations:
- CYP
- cytochrome P450
- 12-HETE
- 12-hydroxy-5,8,10,14-eicosatrienoic acid
- 12-HETrE
- 12-hydroxy-5,8,14-eicosatrienoic acid
- 17-ODYA
- 17-octadecynoic acid
- CDC
- cinnamyl-3,4-dihydroxy-α-cyanocinnamate
- NCI-GC/MS
- negative chemical ionization-gas chromatography/mass spectroscopy
- RP-HPLC
- reverse-phase high performance liquid chromatography
- DMEM
- Dulbecco’s modified Eagle’s medium
- PMSF
- phenylmethanesulfonylfluoride
- PFB
- pentafluorobenzyl ester
- TMS
- trimethylsilyl ether
- LH
- lactalbumin enzymatic hydrolysate
- FBS
- fetal bovine serum
- IL
- interleukin
- GC/MS
- gas chromatography/mass spectroscopy
- LPS
- lipopolysaccharide
-
- Received March 23, 1998.
- Accepted June 23, 1998.
- The American Society for Pharmacology and Experimental Therapeutics
