Detection and Subcellular Localization of Two 15S-Lipoxygenases in Human Cornea

Abstract

purpose. There are two human 15-lipoxygenases (LOX), 15-LOX-1 and -2, which convert arachidonic acid to 15S-hydroxyeicosatetraenoic acid (15S-HETE). The presence of both 15-LOXs in the human cornea prompted this study to delineate their roles in the human corneal epithelium.

methods. Human corneal epithelia from donor corneas and a human corneal epithelial (HCE) cell line were used in [1-¹⁴C]arachidonic acid incubations, Western blot analysis, and quantitative real-time RT-PCR. Cell cultures of HCE were treated with 15S-HETE to measure its effect on cell growth. HCE cells were transfected with plasmids to express green fluorescent (GFP) fusion proteins of 15-LOX-1 and -2, and in vivo laser confocal microscopy was performed to determine the subcellular localization of the 15-LOX fusion proteins.

results. [1-¹⁴C]Arachidonic acid incubations yielded 15S-HETE as the only LOX product. Treatment with 15S-HETE (5–10 μM) reduced growth rate and induced apoptosis in cultured HCE cells in a dose-dependent manner. 15-LOX-2 but not 15-LOX-1 was detected by Western blot analysis, although we were able to detect similar levels of both 15-LOX mRNAs by real-time quantitative RT-PCR. 15-LOX-1 and -2 proteins showed different subcellular expression patterns. 15-LOX-2 GFP was expressed in the cytoplasm and nucleus (actively taken up into the nucleus). 15-LOX-1 GFP fusion protein expression was restricted to the cytoplasm.

conclusions. These findings indicate that 15-LOX-2 is the predominant 15-LOX protein in human cornea, and its product, 15S-HETE, plays a role in cellular proliferation. Because the two 15-LOXs have different subcellular compartmentalization, the authors hypothesize that their products are also compartmentalized and therefore exert different molecular effects in the human corneal epithelium.

Lipoxygenases (LOXs) constitute a major pathway for the conversion of arachidonic acid and other polyunsaturated fatty acids into lipid-signaling molecules. LOX adds a molecule of oxygen to the fatty acid substrate to form initially a hydroperoxide that is subsequently reduced within the cell to the corresponding hydroxyeicosatetraenoic acid (HETE). Two other prominent pathways of oxygenation of arachidonic acid are initiated by cyclooxygenases, leading to prostaglandins, and by the P450 monooxygenases, leading to epoxy- and hydroxyarachidonic acid derivatives. 

There are two human 15-LOXs: 15-LOX-1 and -2. 15-LOX-1 was first identified three decades ago in rabbit reticulocytes. ¹ 15-LOX-2 was identified more recently, and its expression was originally reported in the prostate, skin, and cornea. ² Although the two 15-LOXs only share 40% amino acid identity, both convert arachidonic acid to 15S-HETE. There are significant differences in their activities, however, as 15-LOX-2 converts arachidonic acid purely to 15S-HETE, whereas 15-LOX-1 produces 15S-HETE (80%) as the major product, along with a minor portion of 12S-HETE (20%). ^{2 3 4}  

It is becoming evident that some LOXs, in particular 15-LOXs, play important roles in epithelial cell proliferation and differentiation. A major physiologic function of 15-LOX-1 is degradation of organelles in certain cells as they terminally differentiate. ^{5 6} 15-LOX-1 was initially reported to participate in the maturation of red blood cells by catalyzing the peroxidation and subsequent degradation of organelle phospholipid membranes. ¹ More recently, van Leyen et al. ⁷ have demonstrated that 15-LOX-1 plays a role in the terminal differentiation of central fiber cells in the lens. As for 15-LOX-2, there is a growing body of evidence that it plays a significant role in cellular proliferation. 15-LOX-2 is normally expressed in the prostate, but its expression is markedly reduced or lost during malignant transformation. ^{8 9 10} Also, recent studies have demonstrated that induction of 15-LOX-2 in prostate cancer cell culture leads to decreased cellular proliferation pointing to 15-LOX-2 functioning as a negative regulator of the cell cycle. ¹⁰  

Proliferation and differentiation of the corneal epithelium is a highly regulated process that maintains the physical requirements the cornea needs to refract light. We have shown the expression of 15-LOX-2 mRNA in human corneal epithelium. ² Before the identification of 15-LOX-2, Oliw ¹¹ identified the expression of a 15-LOX by immunohistochemistry in all layers of the human corneal epithelium. However, little is known regarding the role of 15-LOX in the human cornea. This study was initiated to analyze the expression of 15-LOXs and their possible involvement in the proliferation and differentiation of human cornea. We provide evidence indicating that both 15-LOXs are expressed in the human cornea with 15-LOX-2 as the predominant isoform. We also establish different subcellular localization of the two 15-LOXs, suggesting distinct physiological roles. 

Materials and Methods

Tissues and Cells

Human corneal tissues were generously provided by a local eye bank. Corneas were harvested and stored at 4°C in preservative (Optisol; Chiron Vision, Irvine, CA). Corneas not placed for clinical use after 10 days were released for this study and processed either for RNA extraction or arachidonic acid metabolism studies. We have found that we could routinely extract nondegraded RNA from these corneas for up to 3 weeks. 

As SV-40 immortalized human corneal epithelial (HCE) cell line was a generous gift from Kaoru Araki-Sasaki. ¹² The cells were cultured in serum-free medium (Keratinocyte-SFM; Invitrogen-Gibco, Grand Island, NY) at 37°C under 5% CO₂. 

[1-¹⁴C]Arachidonic Acid Incubation and HPLC Analyses

Epithelia from donor corneas were isolated by scraping with a scalpel and placed in incubation buffer (50 mM Tris and 150 mM NaCl [pH 7.5]). HCE cells were cultured to full confluence, collected with a cell scraper, and placed in incubation buffer. A sonicator was used to disrupt the epithelium. Incubations with [1-¹⁴C]arachidonic acid (50 μM) were performed with epithelial homogenate (1 μg/μL protein concentration) at 37°C for 60 minutes. Incubations were stopped by addition of cold methanol (2.5× volume) followed by vigorous vortexing and addition of methylene chloride (1.25× volume). The precipitated protein was removed by centrifugation (10,000g × 2 minutes), and the supernatant was evaporated under nitrogen. The sample was resuspended in 50 μL methanol and triphenylphosphine (1 μg) was added to reduce hydroperoxides to hydroxy products (HPETE to HETE). The products were extracted using a C-18 Bond Elute cartridge (Varian, Inc., Sunnyvale, CA). One third of the sample was analyzed by reversed-phase HPLC with an online radioactivity monitor (Flo-One; Radiomatic Instrument, Tampa, FL) as previously described using a solvent system of methanol-water-acetic acid (80:20:0.01) at a flow rate of 1 mL/min. ² Cold HETE standards were added to monitor the retention times. 

Western Blot Analyses

Western blot analysis was performed with 20 μg of total protein of the human corneal epithelium and HCE cells on SDS-polyacrylamide gels. Standards to both 15-LOX proteins and the polyclonal antibody against 15-LOX-2 were prepared as previously described. ⁹ The proteins were transferred electrophoretically to nitrocellulose. Subsequently, the same blot was stripped and probed with an antibody to 15-LOX-1 purchased from Cayman Chemical (Ann Arbor, MI). We also used an antibody to 15-LOX-1, a gift from Joseph Cornicelli (Parke-Davis/Pfizer, Morris Plains, NJ). Specifically bound protein was detected by a chemiluminescence method (ECL) and exposure to autoradiograph film (Hyperfilm; Amersham Life Science). 

HETE Treatment of HCE Cells

Growth Curves

Growth curves for HCE were established at a seeding density of 5000 cells/cm². At days 4 and 8, changes in cell density were determined by resuspending the adherent cells with trypsin and counting the total cell number using a particle counter (model Z1; Beckman Coulter, Hialeah, FL) according to the manufacturer’s instructions. Cells were also treated with 15S- and 11S-HETE. 11S-HETE was used as the control, since it has not been reported to be a product of arachidonic acid conversion by corneal epithelium. 15S-HETE was added at 5- and 10-μM concentrations. 11S-HETE was added at a 10-μM concentration. HETEs were added to the incubation medium from a stock solution in ethanol (final concentration, <0.05%). 

Our preliminary experiments had shown that both 15S- and 11S-HETE at 10 μM inhibited growth of the HCE and even killed the cells with similar efficacy. These initial experiments were done changing the HETE-containing growth medium only every 3 to 4 days, and the medium was stored premixed with the HETEs. It appears that under these conditions the HETEs may have been converted to toxic metabolites that were responsible for the observed effects. Recent studies have shown that hydroxy fatty acid metabolites can undergo efficient transformation into cytotoxic metabolites under autoxidative conditions. ¹³ Based on these initial findings we modified the assay conditions and added the HETEs only immediately before use to the culture medium, and the medium was changed every 2 days. After this procedure, autoxidative transformation of the HETEs was minimal, and the observed effects were specific to 15S-HETE. 

BrdU and TUNEL Assays

A bromodeoxyuridine (BrdU) uptake assay kit (In-Situ Detection Kit; BD Pharmingen, San Diego, CA) was used to determine whether the cell cycle was altered by HETE treatment. Cells were cultured on glass chamber slides and treated with 11S and 15S-HETE as described earlier. At day 4, cells were pulsed for 1 hour with BrdU. The percentage of cells with BrdU incorporation was determined by counting cells per high-power field. Six fields were counted for each sample. 

A TUNEL assay was performed with a commercially available kit (Fluorescein In Situ Cell Death Detection Kit; Roche Molecular Biochemicals, Indianapolis, IN). Cell culture and treatment conditions are as described for the growth curve assay. At day 4, adherent cells were detached with trypsin and combined with cells floating in the incubation medium. The percentage of cells positive for the TUNEL reaction was determined by flow cytometry (FACSCaliber; BD Biosciences, Franklin Lakes, NJ). Ratiometric analysis was performed on computer (WinList ver. 5.0; Verity Software House, Inc., Topsham, ME) to determine the number of TUNEL-positive cells. ¹⁴  

Real-Time Quantitative RT-PCR

Total RNA was extracted from human corneal epithelium and cell culture samples (Tri-reagent; Molecular Research, Cincinnati, OH) according to the manufacturer’s instructions. HCE cells were harvested when cultures reached full confluence. mRNA copy numbers for 15-LOX-2, 15-LOX-1, and β-actin were determined by quantitative real-time RT-PCR using a fluorescence temperature rapid-air cycler (Lightcycler; Roche Molecular Biochemicals) and probe (SYBR Green; Roche Molecular Biochemicals) as described elsewhere. ¹⁵ mRNA copy numbers of 15-LOX-2 and -1 were normalized to the mRNA copy number for β-actin, which was determined on the same RNA aliquots. Amplifications were performed in glass capillary tubes, using a 20-μL reaction of the following: 200 ng total RNA, 5 mM MgCl₂, 2.0 μL RNA master mix (SYBR Green; Roche Molecular Biochemicals), 0.4 μL reverse transcriptase enzyme (Roche Molecular Biochemicals), and 1.0 μM of each primer. Standard curves were generated for each assay run. The template for the 15-LOX-2 standard curve was a 1088-bp fragment amplified from genomic DNA with the primers 5′-TGC-CTC-TCG-CCA-TCC-AGC-T-3′ and 5′-TGG-GAT-GTC-ATC-TGG-GCC-TGT-3′. The purified product was inserted into the pCRII vector (TA Cloning Kit; Invitrogen, Carlsbad, CA). In the one-step RT-PCR reaction, standard curves with this template were essentially identical with those generated with in vitro transcribed RNA templates. ¹⁶ The template for 15-LOX-1 was a 1027-bp cDNA amplified from genomic DNA using the primers 5′-TGG-TCA-TCC-AGC-TCC-AGC-T-3′ and 5′-ACC-TGT-ACG-GGA-TGC-TCA-3′ and inserted into pCRII. The templates for β-actin and GAPDH were as described previously. ¹⁷  

All primers for real-time assays spanned at least one intron, such that any contaminating DNA would not contribute to the amplimer on which quantification is based. The primers for 15-LOX-2 were 5′-TGC-CTC-TCG-CCA-TCC-AGC-T-3′ (forward) and 5′-TCA-TGG-AAG-GAG-AAC-TCG-GCA-T-3′ (reverse), which give a 126-bp amplified product. The primers for 15-LOX-1 were 5′-GCG-CTG-CGG-CTC-TGG-GAA-ATC-3′ (forward) and 5′-AGA-AGT-GGA-CGT-GGC-CGG-TTG-TG-3′ (reverse), which give a 231-bp product. The primers and real-time quantitative RT-PCR for β-actin and 15-LOX-2 were those reported previously. ¹⁷ The one-step real-time RT-PCR reactions consisted of the following steps: reverse transcription at 55°C for 15 minutes, denaturation at 95°C 30 seconds, amplification for either 45 cycles, and melting-curve analysis from 75°C at a rate of 0.1°C per second under continuous fluorescence monitoring. The amplification programs consisted of heating at 20°C per second to 95°C; cooling at 20°C per second to 55°C; annealing at 55°C for 5 seconds; heating at 20°C per second to 72°C; elongation at 72°C for 11 seconds for 15-LOX-1, 20 seconds for 15-LOX-2, and 19 seconds for β-actin; and heating at either 2°C or 5°C per second to 87°C (β-actin, 15-LOX-1) or 88°C (15-LOX-2), for fluorescence acquisition. The specificity of the amplimer in each reaction was confirmed by the melting curve analysis, with initial gel confirmation that this large peak corresponded to the expected amplimer. The contribution to fluorescence signal of any nonspecific products and/or primer dimers was eliminated by increasing the temperature to 2°C below the melting temperature of the specific product, which eliminated any other minor cDNAs (which have lower melting temperatures). The copy number of mRNA was calculated from serially diluted standard curves generated from the purified cDNA template. Serial dilutions (1:10) over a range of five to six orders of magnitude (e.g., 10⁹–10⁴ copies for 15-LOX-1) were used to generate the standard curves. For each assay, the serially diluted standards were simultaneously amplified with the unknown samples to generate a linear standard curve using the fit points method of analysis with four points or the software-determined second-derivative maximum method. Standard curves for β-actin, 15-LOX-2, and 15-LOX-1 all had correlation coefficients of 0.99 or 1.00 in each assay. Control samples run in triplicate had a variance of approximately 10%. 

Subcellular Localization of 15-LOXs by Laser Confocal Microscopy

The 15-LOX-1 and -2 cDNAs were cloned into the multiple cloning site of the pEGFP vector (Clontech) using the XhoI and EcoRI restriction sites. The green fluorescent protein (GFP) was placed at the N terminus to minimize interference with LOX enzymatic activity, since the C-terminal amino acid serves as a ligand to the active site iron in LOX enzymes. Sequencing was performed to ensure insertion of the cDNAs in the correct reading frame. Arachidonic acid incubation revealed that the LOX enzymatic activity for the GFP fusion proteins was two- to threefold lower compared with the untagged LOX proteins. HCE cells were cultured in 35-mm dishes with a coverslip bottom (Martek, Columbia, MD). When cells reached approximately 30% confluence, transient transfections were performed with the 15-LOX GFP and GFP plasmids using transfection reagents (Lipofectamine and Plus; Invitrogen) according to manufacturer’s instructions. We found the transfection efficiency on HCE to be extremely low (<1%) if the cells approached full confluence. However, we were able to achieve a transfection efficiency of approximately 3% to 5% if cells were <50% confluent. Thirty-six hours after transfection, subcellular localization of 15-LOXs expression was determined on live cells with an inverted laser scanning confocal microscope (model LSM 510; Carl Zeiss Meditec). Laser scans were accomplished with an Argon laser tuned to 488 nm and laser power <10%. 

For fluorescence recovery after photobleaching (FRAP) experiments, the photobleaching component of the microscope software was used to photobleach GFP selectively in either the nucleus or cytoplasm. This was accomplished by selecting the region of interest, and GFP was photobleached with 100% laser power for between 10 and 20 scans. The photobleaching resulted in reducing emission intensity by between 90% and 95%. In some experiments, fluorescence recovery was allowed to occur over time. Fluorescence recovery was demonstrated by determining the change in the fluorescence emission intensity in the bleached region of interest over time. 

Results

Arachidonic Acid Metabolism by Corneal Epithelial Tissue and Cultured HCE Cells

The metabolism of [1-¹⁴C]arachidonic acid in the corneas of human donors was analyzed with reversed phase (RP)-HPLC (Fig. 1) . The RP-HPLC system used for product analysis is capable of resolving all possible LOX-derived HETE isomers. The main [¹⁴C]-labeled arachidonic acid product of the incubations cochromatographed with an authentic standard of 15S-HETE. The 15-HETE product was isolated and further analyzed using chiral phase HPLC to analyze the configuration of the 15-hydroxy group. Chiral analysis revealed that 15-HETE formed in the corneal epithelium was exclusively of the 15S configuration (not shown), providing evidence for its formation by a LOX. Further analysis of 10 additional samples from different donors revealed a high degree of variation in the level of 15S-HETE formation when normalized to protein content. Most samples converted approximately 20% to 30% of the substrate, although we detected between <5% and ≤80% conversion of arachidonic acid in individual samples. In addition to forming 15-HETE, some of the samples showed formation of a distinct polar product, which was probably due to cyclooxygenase activity. 

Metabolism of [1-¹⁴C]arachidonic acid was also analyzed in an SV40-immortalized cell line derived from human corneal epithelium (HCE). As expected, the corneal cells also formed 15S-HETE as the only LOX product (results not shown). In the cultured corneal epithelial cells, the LOX activity was lower than in the corneal tissue when normalized to a similar amount of protein. 

The incubations with radiolabeled arachidonic acid also revealed that activity of other LOX isozymes (5-LOX and 12-LOX) was absent in the corneal tissue and in HCE cells. Furthermore, P450 metabolism of arachidonic acid usually yields a mixture of different epoxyeicosatrienoic acid isomers and HETEs, including the ω- and ω-1–hydroxy products 20-HETE and 19-HETE. ^{18 19} We detected formation of only one major HETE enantiomer, 15S-HETE. We did not detect any ω/ω-1-hydroxy products, which have a shorter retention time compared with 15S-HETE using our RP-HPLC system. Therefore, the contribution of P450 monooxygenases in the formation of HETE products could be excluded based on the HPLC analyses. Prompted by reports on the identification of low concentrations of 12-HETE in human tear film, ²⁰ we carefully analyzed for the formation of this product but were unable to detect 12-HETE in any of the samples. 

Detection of 15-LOXs by Western Analyses and Real-Time Quantitative RT-PCR

The results of the arachidonic acid metabolism studies pointed to the presence a 15-LOX in human corneal epithelium and in HCE cells. Because there are two LOXs forming 15-HETE in humans, we analyzed for the presence of 15-LOX-1 and -2 with Western blot and real-time quantitative RT-PCR analyses. 

Western blot analysis was performed with total protein isolated from human corneal epithelium (Fig. 2) . Expression of 15-LOX-2 was detected in all three corneal samples, although we noticed a variation in the abundance of 15-LOX-2 protein expression in different samples. The corneas obtained had been stored for 10 to 21 days at 4°C in preservative medium (Optisol; Chiron), implying that extended storage had no effect on 15-LOX-2 detection. Expression of 15-LOX-1 was not detected in the Western blot analysis of the same samples, even though we had a positive control of 15-LOX-1 protein detectable at the level of 5 ng (not shown). We also analyzed snap-frozen samples from freshly harvested eyes to determine whether 15-LOX-1 is degraded with extended storage. We obtained identical results with snap-frozen tissues as in the eye bank corneas. Western analyses of the HCE cells failed to detect the presence of 15-LOX-1 or -2, although we were able to detect low levels of 15-HETE formation. We conclude that the protein levels of both LOXs in HCE cells and the level of 15-LOX-1 in the donor corneas were below the sensitivity of our Westerns analyses. 

We used quantitative real-time RT-PCR to analyze for the expression of 15-LOX-1 and -2 mRNA in human corneal epithelium and in HCE cells (Table 1) . The results were standardized to β-actin mRNA copy numbers. In four human corneal samples, we found approximately a 10-fold variation in the copy number of both LOX mRNAs, with no identifiable trends to indicate which 15-LOX had the higher mRNA copy number. In HCE cells, by contrast, we found 15-LOX-2 message level to be seven times greater than 15-LOX-1 mRNA levels. The levels of 15-LOXs mRNA were several orders of magnitude lower in HCE cells than in donor corneas, a trend similar to the 15-LOX enzymatic activity as measured by 15S-HETE formation. 

Effect of 15S-HETE on Cellular Proliferation of HCE

To gain insight into the potential function of 15-LOX, we analyzed the growth of HCE cells in the presence of 15S-HETE. For controls, we used no treatment and 11S-HETE, an isomeric arachidonic acid metabolite. Growth curves were established for treatment with 5 and 10 μM 15S-HETE and 11S-HETE (Fig. 3) . 11S-HETE did not significantly alter the growth curve, and both no treatment and 11S-HETE groups exhibited similar doubling times of 2.6 and 2.7 days, respectively. In contrast, treatment with 15S-HETE significantly decreased the number of cells at days 4 and 8 in a dose-dependent manner. The doubling time was prolonged to 3.3 days with 5 μM 15S-HETE treatment. At 10 μM 15S-HETE, there was a decline in the number of cells at each time point, resulting in a calculated halving time of 12.7 days. 

BrdU and TUNEL experiments were performed to determine whether the observed reduction in cell number with 15S-HETE was due to disruption of cellular proliferation or increased cell death. In BrdU assays, there was no difference between the treatment and control groups (Fig. 4A) , indicating that the percentage of cells in S phase was not affected by 15S-HETE treatment. However, there was a higher percentage of cells positive for the TUNEL reaction with 15S-HETE treatment (Fig. 4B)in a dose-dependent fashion. At day 4, approximately 2% of the cells were positive for the TUNEL reaction in both no treatment and the 11S-HETE treatment group, indicating the rate of apoptosis in these two groups are similar. With 5 and 10 μM 15S-HETE treatment, the percentage of cells positive for TUNEL increased to 7% and 13%, respectively. These results indicate that increased cell death rather than decreased cellular proliferation accounts for the observed differences in growth curves between 15S-HETE treatment and control groups. 

Subcellular Localization of 15-LOX Proteins

For the localization experiments, we used transient transfection of LOX GFP constructs into HCE cells. The different subcellular localization patterns of the two 15-LOX GFP fusion proteins was readily apparent from in vivo laser confocal microscopy (Fig. 5) . 15-LOX-1 GFP expression was only detected in the cytoplasm with no fluorescence in the nucleus. In contrast, 15-LOX-2 GFP expression was detected in both the cytoplasm and the nucleus with similar fluorescence intensity. 

Because in our initial analysis it appeared that 15-LOX-2 GFP was expressed in a similar pattern as GFP alone (Fig. 6A , inset), we used FRAP experiments to analyze subcellular trafficking of the protein. In FRAP, the laser of the confocal microscope is targeted at a fixed region restricted to the cytoplasm (Fig. 6A , arrowheads). Photobleaching in cells expressing GFP led to a decrease in nuclear and cytoplasmic fluorescence (Fig. 6B , inset) indicating rapid mobility of GFP molecules between nuclear and cytoplasmic compartments of the HCE cells. In 15-LOX-2 GFP expressing cells, however, the decrease in fluorescence intensity was observed only in the cytoplasm (Fig. 6B , arrowhead and asterisk), indicating restricted movement of 15-LOX-2 between the nuclear and cytoplasmic cellular compartments. 

Discussion

In this study, we have demonstrated that 15S-HETE is the major LOX product formed by the human corneal epithelium.15-LOX-1 and -2 mRNA were expressed in similar abundance. However, at the protein level only 15-LOX-2 expression was detected by Western blot analysis, and, as the Western analyses had similar sensitivities for the two proteins, we conclude that 15-LOX-2 is the predominant constitutively active LOX expressed in the human corneal epithelium. This conclusion is also supported by analysis of the product formation. Pure 15S-HETE was detected, implying that 15-LOX-2 is the major active 15-LOX in the human corneal epithelium. In contrast, 15-LOX-1 converted arachidonic acid to 15-HETE and 12-HETE at a ratio of approximately 4 to 1, and formation of 12-HETE was not detected in our HPLC analyses. 

Although Liminga et al. ²¹ have detected the localization of 15-LOX-1 protein in all layers of the human corneal epithelium by using immunohistochemistry, we were unable to detect 15-LOX-1 in our Western analyses. A possible explanation for the lack of detection of 15-LOX-1 protein in light of similar message levels of both 15-LOXs is that 15-LOX-1 mRNA is silenced in human corneal epithelium by a mechanism similar to that reported in erythroid cells, ^{22 23} in which, 15-LOX-1 message is accumulated and silenced by a regulatory element at the 3′ untranslated region. The abundance of the 15-LOX-1 message in erythroid precursor cells is second only to mRNA for hemoglobin. ^{23 24} The organelles are degraded during the terminal differentiation of red blood cells. As these cells mature to red blood cells, 15-LOX-1 message is unsilenced, leading to an increase in the 15-LOX-1 protein and degradation of organelles. ^{7 23 24 25} It is possible that 15-LOX-1 protein in human corneal epithelium is also temporally regulated. The expression in human cornea may be limited to a narrow time window, as reported in lens fiber cell differentiation. ⁷ This could account for our lack of detection of the corresponding protein. 

The human cornea is distinctive in that only 15-HETE formation is reported as the sole LOX product. ^{26 27} Typically, 12-HETE is the most frequently reported HETE in the mammalian cornea. Bazan et al., ^{28 29} Birkle et al., ³⁰ and Ottino et al. ³¹ have extensively reported on 12-LOX expression in rabbits and have demonstrated 12-lipoxgenase to be involved in wound healing and the inflammatory response. ^{28 29 30 31} It appears that different species use different LOX and/or different LOX metabolites, and therefore it is difficult to apply results obtained from animal LOX studies to the human cornea. 

The 15-LOX product, 15S-HETE, appears to induce apoptosis in HCE cells in a dose-dependent fashion. Tang et al. ¹⁰ have also reported similar effects of 15S-HETE in various prostate cell lines. They report that the addition of exogenous 15S-HETE also led to an increased rate of apoptosis in a dose-dependent manner, but at concentrations >25 μM. At lower concentrations, cell cycle progression was disrupted as indicated by decreased BrdU incorporation. We did not observe any apparent changes in the rate of BrdU incorporation in HCE cells with 15S-HETE treatment, probably reflecting a difference in sensitivity in our cell line. Exogenous 15S-HETE at 5 μM is sufficient to induce apoptosis in HCE cells whereas prostate cells require much higher levels. Nevertheless, the observed overall effect of exogenous 15S-HETE treatment is decreased cell number and increased cellular death, secondary to apoptosis. 

The presence of the two 15-LOXs in human corneal epithelium raises the question of why there are two enzymes capable of producing the same product, 15S-HETE. Because 15S-HETE is a product of both 15-LOXs, we are unable to directly discriminate the different physiologic roles of the two 15-LOXs expressed in the HCE cell by using exogenous 15S-HETE. Results from studies on the subcellular localization of the two 15-LOXs using laser confocal microscopy suggest different physiologic roles in the human corneal epithelium. Expression of 15-LOX-1 GFP was observed strictly in the cytoplasm, while 15-LOX-2 GFP was observed in the cytoplasm and in the nucleus. In our FRAP experiments, the mobility of 15-LOX-2 between the nuclear and cytoplasmic compartment is limited, indicating that 15-LOX-2 is actively accumulated in the nucleus. A putative bipartite nuclear localization sequence has been identified for 15-LOX-2. ⁸ The subcellular distribution of 15-LOX-2 in both the cytoplasm and the nucleus has been reported in prostate cell lines, ⁸ similar to our findings in HCE cells. Also, in immunohistochemistry analysis of 15-LOX-1 on human cornea by Liminga et al., ²⁷ 15-LOX-1 in the basal layer of the human cornea can be seen localized primarily to the cytoplasm. Because the two LOXs have different compartmentalization within the cell, it is reasonable to expect they will have access to different compartmentalized pools of fatty acid substrates, and their products will also be compartmentalized. It is also important to note that both 15-LOXs are active with linoleic acid as a substrate, forming 13S-hydroxyoctadecadienoic acid (13S-HODE) as a product. In fact, in vitro, linoleic acid is oxygenated by 15-LOX-1 with approximately 50% higher catalytic efficiency than arachidonic acid, ³² leading to the speculation that linoleic acid could be the more physiologic substrate for this enzyme. ³³ In contrast, 15-LOX-2 appears to preferentially metabolize arachidonic acid over linoleic acid. ² Therefore, the two 15-LOXs may exert different molecular effects through different cellular compartmentalization and substrate specificity. 

The physiologic roles of the two 15-LOXs are continuing to be better delineated, with evidence pointing to regulation of cellular proliferation by 15-LOX-2 and terminal cellular differentiation by 15-LOX-1. However, the presence and physiologic roles of these two proteins within the same cell is not well defined. In this report, we provide evidence that (1) mRNA of 15-LOX-1 and -2 are expressed in the human corneal epithelium, (2) 15-LOX-2 is the predominant active LOX in the human corneal epithelium, and (3) both 15-LOXs have different subcellular expression patterns when transfected into HCE cells. In the human corneal epithelium, cellular proliferation, and differentiation is highly regulated. To maintain the cornea’s optical properties, the number of cells entering the corneal epithelium from the limbus must equal the number of terminally differentiated cells lost at the epithelial surface. The interplay of 15-LOX-2 and -1 expression levels and localization may play an important role in the maintenance of the human corneal epithelium through regulation of corneal cellular proliferation and differentiation. 

 Supported by National Eye Institute Grants EY13592 and P30 EY08126, National Institute of Medical Sciences Grant GM53638, and a challenge grant from Research to Prevent Blindness Inc.

 Submitted for publication October 1, 2004; revised November 17, 2004; accepted November 20, 2004.

 Disclosure: M.S. Chang, None; C. Schneider, None; R.L. Roberts, None; S.B. Shappell, None; F.R. Haselton, None; W.E. Boeglin, None; A.R. Brash, 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.

 

Arachidonic acid metabolism by human corneal epithelium. Incubation of [¹⁴C]-arachidonic acid with corneal homogenates was performed and analyzed by RP-HPLC. A prominent [¹⁴C]15S-HETE peak was seen on the chromatogram eluting at ≈14 minutes. The small peak just before the 15S-HETE was provisionally identified as 15-keto-eicosatetraenoic acid, on the basis of retention time. Retention times of unlabeled HETEs are indicated on the chromatogram.

Figure 2.

 

Western blot detection of 15-LOX-2 in human corneal epithelium. Epithelia (20 μg total protein) from three human corneas were analyzed for 15-LOX-2 expression. Corneas 1, 2, and 3 were stored in preservative (4 C) for 10, 15, and 21 days, respectively. Benign human prostate, purified 15-LOX-2 (5 ng), and purified 15-LOX-1 (5 ng) protein were included as the controls. Western blot analysis was performed with a rabbit polyclonal antibody and chemiluminescence detection. A band was visible at ≈80 kDa in the human cornea, human prostate, and purified 15-LOX-2 standard. No band was detected in the 15-LOX-1 lane, indicating no cross reactivity of the 15-LOX-2 antibody.

Table 1.

 

Determination of 15-LOX-1 and 15-LOX-2 mRNA by Real-Time RT-PCR

Sample	15-LOX-1/β-Actin	15-LOX-2/β-Actin
Human cornea 1	2.56 × 10−3	7.59 × 10−3
Human cornea 2	1.02 × 10−3	4.46 × 10−3
Human cornea 3	9.97 × 10−3	1.61 × 10−3
Human cornea 4	2.02 × 10−3	7.42 × 10−4
HCE cells (avg. ± SD, n = 4)	4.73 × 10−6 (±9.5 × 10−7)	3.5 × 10−5 (±7.6 × 10−6)

 The relative abundance of 15-LOX-1 and 15-LOX-2 mRNA were determined from four human corneal epithelia and confluent HCE cell cultures (n = 4). The levels of 15-LOXs were standardized to β-actin mRNA levels. The relative abundance of 15-LOX-1 and 15-LOX-2 mRNA exhibited almost a 10-fold variation, with no trends indicating which 15-LOX was more highly expressed at mRNA level. The HCE cell culture demonstrated a significantly greater relative abundance (P < 0.05) for 15-LOX-2 mRNA by seven-fold over 15-LOX-1.

Figure 3.

 

Growth curve of human corneal epithelial cells. HCE cells immortalized by SV-40 antigen were seeded at 5000 cells/cm². Cells were treated with 15S-HETE (5 and 10 μM) and 11S-HETE (10 μM). The culture cell densities at 4 and 8 days were compared with densities of untreated cells (n = 4 with SEM). Cell cultures treated with 15S-HETE revealed significantly lower numbers of cells compared with untreated cultures (*P < 0.05, by Student’s t-test). Treatment with 11RS-HETE did not significantly alter the number of cells.

Figure 4.

 

BrdU incorporation and TUNEL reaction assay. HCE cell cultures were treated with 15S-HETE (5 and 10 μM) and 11S-HETE for 4 days. (A) Cell cultures were pulsed with BrdU for 1 hour, and the results were determined by counting cells positive for BrdU uptake per high-power field (six fields per culture, n = 5 with SEM). There was no significant difference in BrdU uptake between the treatment and no-treatment groups. (B) A TUNEL reaction was performed with nonadherent and trypsinized cells. The TUNEL-positive cells were detected by flow cytometry (n = 5 with SEM). 11S-HETE and no treatment groups have similar percentages of positive cells. Treatment with 15S-HETE led to a dose-dependent and significant increase (*P < 0.05) in the percentage of TUNEL-positive cells, indicating 15S-HETE treatment led to increased cell death by apoptosis.

Figure 5.

 

In vivo subcellular localization of 15-LOX in human corneal epithelial cells by laser confocal microscopy. HCE cells were transfected with plasmids to express GFP fusion proteins of 15-LOX-1 (A) and 15-LOX-2 (B). Subcellular localization of 15-LOX-1 GFP (A) is restricted to the cytoplasm. (B) 15-LOX-2 GFP expression was detected in both the nucleus and the cytoplasm.

Figure 6.

 

FRAP assay with HCE cells expressing fluorescent proteins. HCE cells expressing a 15-LOX-2 GFP fusion protein or GFP alone (inset) were analyzed in vivo using laser confocal microscopy. (A) Nuclear and cytoplasmic localization is seen for both proteins with similar fluorescence intensity in the cytoplasm and nucleus. The area of photobleaching is restricted to a portion of the cytoplasm (arrowheads). The cytoplasm of the cell on the left ( ) had been photobleached 5 minutes earlier, and this cell demonstrated greater fluorescence in the nucleus than in the cytoplasm. (B) Fluorescence recovery was allowed for 5 minutes, and cells were reanalyzed by laser confocal microscopy. In the cell expressing GFP alone (inset) the fluorescence was barely visible, demonstrating an equivalent decrease in nuclear and cytoplasmic fluorescence. Cells expressing 15-LOX-2 GFP showed only decreased cytoplasmic fluorescence. The nuclear fluorescence intensity continued to be higher than the cytoplasmic even after 10 minutes ( ). These results indicate that nuclear 15-LOX-2 is protected from photobleaching targeted at the cytoplasm and that there is restricted exchange of 15-LOX-2 between the nucleus and cytoplasm. The bright band of fluorescent dots below the left cell is an artifact.

The authors thank Sai Han Presley for technical assistance; Peggy Hall and Craig Henderson of Tennessee Donor Services for their generous assistance in providing the eye bank corneas for the study; and Christopher J. Pino for assistance with flow cytometry.