A Biosynthetic Pathway Generating 12-Hydroxy-5,8,14-eicosatrienoic Acid from Arachidonic Acid Is Active in Mouse Skin Microsomes

Cyclooxygenases, lipoxygenases, and cytochromes P450 (P450 gene products) activate arachidonic acid. Prostanoids, hydroxy and epoxy fatty acids generated by these enzyme systems, mediate signals involved in epidermal homeostasis, skin diseases, and cancers. These lipid mediators are implicated in mechanisms regulating diverse cellular processes including epidermal cell proliferation and differentiation, cutaneous responses to inflammation and environmental cell damage, and cutaneous immune responses (for review, see Ikai, 1999; Iversen and Kragballe, 2000; Marks et al., 2000; Ziboh et al., 2000; Kabashima and Miyachi, 2004).

In mouse skin, epidermal cyclooxygenases generate mainly prostaglandins E₂ and D₂ from arachidonate. 12-Hydroxyeicosatetraenoic acid (12-HETE) is the main hydroxy fatty acid generated (Hammarström et al., 1979; Bedford et al., 1983; Ruzicka et al., 1983), and levels of these enzyme activities are regulated as a function of the differentiation state of the epidermal cells (Henke et al., 1986; Cameron et al., 1990). In human skin, epidermal cyclooxygenases generate mainly prostaglandins E₂ and F_2α from arachidonate. The main hydroxy fatty acids generated are 12- and 15-HETE, and the ratio of these regioisomers is influenced by the differentiation state of epidermal cells (for review, see Iversen and Kragballe, 2000; Ziboh et al., 2000).

The cytochrome P450 superfamily includes 57 putatively functional P450 genes in humans (Nelson et al., 2004). P450 enzymes have critical roles in metabolism of diverse natural and foreign compounds (Nebert and Russell, 2002). Several are active toward arachidonate in in vitro studies, especially CYP1, CYP2, CYP3, and CYP4 family members. Arachidonic acid monooxygenases typically generate multiple products in a regio- and stereospecific manner (Capdevila and Falck, 2002). In addition to epoxy- and ω/ω-1 hydroxy fatty acids, they can catalyze bisallylic oxidations such that in a given tissue 12- and 15-HETE regioisomers, for example, potentially arise from both lipoxygenase- and P450-dependent metabolism.

The biosynthesis and functions of P450-derived eicosanoids in the skin are poorly understood even though this enzyme system likely generates a large number of biologically active lipid mediators in epidermal cells. Cytochromes P450 require electron transfer proteins to transfer electrons from NADPH to the enzyme active site. Hence, P450-derived products are generally not observed under assay conditions used to study cyclooxygenase- and lipoxygenase-mediated metabolism. The first evidence that epidermal cells generate P450-derived epoxyeicosatrienoic (EET) acids was obtained using human keratinocyte cell suspensions, in which cell membranes and reducing environment remained intact (Holtzman et al., 1989). Although analyte identity was not proven in these studies, mass spectrometry was used in a subsequent study to measure cellular levels of EETs arising from endogenous arachidonic acid in both human and mouse epidermal cells (Ladd et al., 2003; Du et al., 2005). These results confirmed that catalytically active epoxygenases are expressed in epidermal cells since P450 enzymes are the only known catalytic source of EETs.

Previously, we identified CYP2B19 as a major source of EET formation in mouse skin (Du et al., 2005). In the present study, we aimed to demonstrate that mouse epidermal microsomes generate EETs when reconstituted in the presence of [1-¹⁴C]arachidonate. To further establish roles for CYP2B19, we aimed to compare the regio- and stereospecificity of EETs formed by mouse epidermal microsomes with those generated by Escherichia coli-expressed CYP2B19 (Keeney et al., 1998). Instead of P450-derived EETs, we identified intermediates in a biosynthetic pathway not previously described in mouse skin that leads to formation of 12-hydroxy-5,8,14-eicosatrienoic acid (12-HETrE).

Materials and Methods

Microsome Preparation. Full-thickness skin or epidermal sheets from newborn mice (CD-1 strain; Charles River Laboratories, Inc., Wilmington, MA) were homogenized in 0.05 M Tris-HCl, pH 7.4, containing 0.25 M sucrose, 0.15 M KCl, and protease inhibitors (Complete-mini; Roche Applied Science, Indianapolis, IN). Epidermal sheets were peeled from the dermis after a 3-h incubation at 4°C in 0.3% Dispase II (Sigma-Aldrich, St. Louis, MO) in phosphate-buffered saline. Tissues were disrupted by three 1-min pluses at maximum speed (≈26,000 rpm), using a Polytron PT3100 fitted with a 12-mm generator with knives (PT-DA3013/2TM; Brinkmann Instruments, Westbury, NY). Samples were chilled for 1 min between pulses. Homogenates were centrifuged for 20 min at 900g and 4°C. The 900g supernatants were centrifuged for 30 min at 12,000g and 4°C. The 12,000g supernatants were centrifuged for 60 min at 100,000g and 4°C. If not used immediately, the 100,000g microsomal pellets were overlaid with 1 ml of 30% glycerol in 0.05 M Tris-HCl, pH 7.4, and frozen at -30°C for up to 2 months. Before use, the glycerol overlay was removed, and pellets were washed with 1 ml of 0.15 M KCl, 0.01 M MgCl₂, and 0.05 M Tris-HCl, pH 7.4.

Arachidonic Acid Metabolism Assays. [1-¹⁴C]Arachidonic acid (48 mCi/mmol; PerkinElmer Life and Analytical Sciences, Boston, MA) was purified by SiO₂ chromatography, and assays were performed at pH 7.4 and 35°C as described previously (Capdevila et al., 1990), with minor modifications. Microsomal pellets were resuspended directly in reaction buffer (0.15 M KCl, 0.01 M MgCl₂, 0.05 M Tris-HCl, pH 7.4, and 2 mg/ml sodium isocitrate) and preincubated at 35°C for 15 min (±inhibitors). [1-¹⁴C]Arachidonate substrate (70–100 μM) was added 2 min prior to starting reactions with NADPH.

The extracted reaction products were reconstituted in initial phase solvent and resolved by reversed-phase HPLC (5-μm Dynamax Microsorb C₁₈ column; 4.6 × 250 mm; Varian, Inc., Palo Alto, CA), using a linear solvent gradient from initially water/acetonitrile/acetic acid (49.95:49.95:0.1; v/v/v) to acetonitrile/acetic acid (99.9:0.1; v/v) for 40 min at 1 ml/min (Capdevila et al., 1990). Effluents were collected separately and resolved by normal-phase HPLC (5-μm Dynamax Microsorb silica column; 4.6 × 250 mm; Varian Inc.), using an isocratic solvent mixture of acetic acid/2-propanol/hexane (0.1:0.5: 99.4; v/v/v) at 1.0 ml/min. Separations were performed using an Alliance Systems 2690 separation module and model 996 photodiode array detector (Waters, Milford, MA). Quantitation of radiolabeled products utilized a β-RAM model 3 radiochromatography detector fitted with a 250-μl lithium glass scintillator solid cell (IN/US Systems Inc., Tampa, FL). Standard curves were generated, correlating picomoles of [1-¹⁴C]arachidonic acid injected versus output signal (microvolts times second). Extraction efficiencies averaged 75%, estimated by addition of radiolabeled internal standards (8,9-/11,12-EET).

Chemical inhibitors and unlabeled eicosanoids used as reference compounds were from Cayman Chemical (Ann Arbor, MI), except for 12-HETrE and 12-oxo-5,8,14-eicosatrienoic acid (12-oxo-ETrE) (from S. G. Jagadeesh and J. R. Falck), N-adamantanyl-N′-dodecanoic acid urea (from C. Morrisseau and B. D. Hammock, University of California, Davis, CA), and 1-aminobenzotriazole (Sigma-Aldrich). The [1-¹⁴C]EET reference standards were synthesized from [1-¹⁴C]arachidonate using 3-chloroperbenzoic acid (Sigma-Aldrich), as described (Falck et al., 1990). Dihydroxyeicosatrienoic acids (DiHETrEs) were prepared from EETs as described (Zeldin et al., 1993) by hydrolysis in 50% CH₃COOH in water for 12 h at 45°C under argon, with constant mixing. Synthesized EETs and DiHETrEs were purified before use by reversed-phase HPLC.

Liquid Chromatography/Mass Spectrometry. LC/MS analyses were performed using a Thermo Electron LCQ DecaXP (Thermo Electron Corporation, Waltham, MA) ion trap mass spectrometer equipped with an Agilent Technologies (Palo Alto, CA) 1100 series binary HPLC pump and thermostated autosampler. Chromatographic separations were done on a 150-× 1.0-mm, 5-μm, 300-Å Jupiter C18 column (Phenomenex, Torrance, CA), using a binary acetonitrile/water gradient at flow rate of 75 μl/min. Solvent A consisted of acetonitrile/water (5:95; v/v) buffered with 10 mM ammonium acetate, pH 5.5. Solvent B consisted of acetonitrile/water (95:5; v/v) buffered with 10 mM ammonium acetate at pH 5.5. The solvent gradient had an initial isocratic hold for 1 min at 35% B, followed by a linear gradient program from 35 to 100% B over 14 min, then held at 100% B for 7 min. LC/MS experiments were done using negative electrospray ionization (ESI) using the standard DecaXP ESI source; nitrogen was used for both the sheath and auxiliary drying gases. The ESI spray potential was held at 3.5 kV, and the heated capillary ion transfer tube was held at 300°C. An ion source declustering potential of 10 eV was employed to enhance signal intensity. Optimal ESI operating conditions were established using a postcolumn infusion of arachidonic acid. Full-scan spectra were recorded over the mass range m/z 200 to 400, at a scan rate of 1.4 spectra/s. Product ion spectra were determined in the ion trap over the mass range m/z 75 to 330 with normalized collision energy of 35%, using data-dependent scanning with a signal intensity threshold of 5 × 10⁵ counts. Products were kept at -20°C until analysis and then held at 5°C in the refrigerated autosampler. Special precautions were taken to prevent isomerization of 12-oxo-ETrE, which was freshly synthesized (by S. G. Jagadeesh and J. R. Falck) the day before analysis and shipped overnight on dry ice to Vanderbilt. Upon arrival, it was immediately diluted in mobile phase and analyzed by LC/MS.

Results

Mouse Epidermal Microsomes Generate NADPH-Dependent Products from Arachidonic Acid. Microsomes were prepared from both the epidermis and full-thickness mouse skin; both were active toward [1-¹⁴C]arachidonic acid and generated similar product profiles. The majority of ¹⁴C products formed in the absence of NADPH, presumably by epidermal lipoxygenases and cyclooxygenases. However, two products (I and II) formed only in the presence of NADPH, potentially arising from cytochrome P450-dependent catalysis (Fig. 1). Product I had the same retention time as the CYP2B19 product 14,15-EET, resolved by reversed-phase HPLC (Fig. 1A). No products were observed when microsomes were preheated at 65°C for 10 min, indicating the two NADPH-dependent products arise enzymatically (data not shown).

We detected product coeluting with a single EET regioisomer (14,15-EET), even though epoxygenases including CYP2B19 typically generate two or more EET regioisomers. This could not be explained by epoxide hydrolase-dependent hydrolysis of EET regioisomers to the corresponding vicinal diols. The soluble epoxide hydrolase inhibitor N-adamantanyl-N′-dodecanoic acid urea (10–100 μM) had no discernible effect on the profile of eicosanoids generated from arachidonate by mouse epidermal microsomes (data not shown). Minor products having similar retention times as authentic DiHETrEs were formed regardless of whether NADPH was added (Fig. 1). There was also no evidence for 5,6-DiHETrE 1,5-lactone, which could be resolved from product I in the reversed-phase HPLC system. We concluded that DiHETrEs were not generated at detectable levels.

Accumulation of the NADPH-dependent product I increased linearly only over 10 min, at protein concentrations up to 1 mg/ml for epidermal microsomes or 4 mg/ml for whole-skin microsomes. We reasoned that NADPH-dependent products would be more easily detected if we could eliminate the competition for substrate by epidermal lipoxygenases and cyclooxygenases. However, we were unable to selectively inhibit these NADPH-independent activities toward arachidonate (inhibitor data not shown). At all concentrations tested, the lipoxygenase inhibitors nordihydroguaiaretic acid (2–50 μM) and baicalein (0.1–5 μM) inhibited at least partially the formation of product I. This effect could be nonspecific or it could indicate lipoxygenase activities were involved in product I formation. Because the cyclooxygenase inhibitor indomethacin (25–50 μM) increased product I formation modestly (range, 30–300%), we routinely preincubated microsomes with 25-μM indomethacin, even though this drug had no measurable effect on NADPH-independent product formation. Inexplicably, the partial inhibition of product I due to nordihydroguaiaretic acid or baicalein was reversed by preincubating microsomes with both indomethacin and lipoxygenase inhibitor.

To prove whether product I was the CYP2B19 product 14,15-EET, a mixture of purified [¹⁴C]-product I and [¹⁴C]-14,15-EET was resolved by normal-phase HPLC. These entities proved to be chemically distinct (Fig. 2A), a result confirmed by the apparent stability of product I in the presence of acetic acid (Fig. 2B). When [1-¹⁴C]-14,15-EET was subjected to the same acidic conditions, [1-¹⁴C]-14,15-DiHETrE formed (Fig. 2C). Previously, we used mass spectrometry to prove that CYP2B19 protein is expressed and that cellular EET accumulates in mouse epidermal keratinocytes (Du et al., 2005). Yet, under the assay conditions, epidermal microsomes did not form detectable levels of EET. It is not clear whether this was due to low P450 protein levels, instability or deleterious effects of sample preparation on P450 activities, low substrate turnover for epoxygenases relative to lipoxygenases and cyclooxygenases, or substrate depletion due to rapid conversion of arachidonate to NADPH-independent products. If product I was not 14,15-EET, what were the identities of the NADPH-dependent products?

Mouse Epidermal Microsomes Generate 12-HETrE from Arachidonic Acid. Spectral data indicated the NADPH-dependent products (I and II) were not conjugated dienes, dienones, or trienes since neither had demonstrable absorbance at 237 or 280 nm (Table 1). Hydroxyeicosatrienoic acid regioisomers were evaluated since these eicosanoids have similar spectral properties as the epidermal products I and II. Neither of the NADPH-dependent products coeluted with 15-hydroxy-8,11,13-eicosatrienoic acid (15-HETrE), resolved by reversed-phase HPLC. Product II coeluted with 12-HETrE (Table 1), an eicosanoid formed from arachidonate by bovine corneal and rat epidermal microsomes (Murphy et al., 1988; Nishimura et al., 1991; Van Wauwe et al., 1991).

To prove whether the intermediates involved in 12-HETrE synthesis were also present in the in vitro reactions, we purified products I to IV (from Fig. 1). Each of these products was mixed individually with a selected reference compound, and the mixtures were resolved on two chromatographic systems. The NADPH-dependent product II was identified as 12-HETrE since it coeluted with this reference compound on reversed-phase and normal-phase HPLC systems (Table 1). Product IV, a major NADPH-independent product, was assumed to be HETE because it had an absorbance maximum at 237 nm, indicative of a conjugated diene chromophore. Product IV coeluted with 12-HETE in both chromatographic systems (Table 1). The minor NADPH-independent product III had an absorbance maximum at 282 nm in our reversed-phase solvent system, indicative of a conjugated dienone chromophore. Product III coeluted with 12-oxo-ETE in both chromatographic systems (Table 1). Finally, the NADPH-dependent product I, previously shown not to be an epoxy fatty acid, coeluted with 12-oxo-ETrE in both chromatographic systems (Table 1). We concluded from these chromatographic and spectral data that mouse epidermal microsomes contain competent enzymes that generate all of the intermediates in the 12-HETrE biosynthetic pathway (Fig. 3).

Mass Spectral Analyses Confirm That Mouse Epidermal Microsomes Generate 12-oxo-ETrE and 12-HETrE from Arachidonic Acid. We confirmed the structures of the two NADPH-dependent products 12-oxo-ETrE (I) and 12-HETrE (II) using LC/ESI/MS (Fig. 4). The LC elution time and ESI tandem mass spectrum of product II were identical to those obtained for authentic 12-HETrE (Fig. 4, A–D). In addition to the carboxylate anion [M-H]^- at m/z 321 and its dehydration product [M-H-H₂O]^- at m/z 303, the two abundant ions m/z 181 and 209 are indicative of hydroxyl position in 12-HETrE/product II. The inset indicates the cleavage and ion structure for these two ions (Fig. 4D). Likewise, Fig. 4, E to H, shows identical LC elution times and ESI tandem mass spectra for product I and authentic 12-oxo-ETrE. Ions at m/z 319, 301, and 275 represent the carboxylate anions [M-H]^-, [M-H-H₂O]^-, and [M-H-CO₂]^-, respectively. Structures for ions m/z 163 and 153 are shown in the inset; ion m/z 163 may result from internal cleavages as indicated (Fig. 4H).

As a positive control, we also analyzed product IV, identified as 12-HETE (Table 1). The LC elution time and ESI tandem mass spectrum of product IV were identical to those for authentic 12-HETE (data not shown). We did not subject product III/12-oxo-ETE to LC/MS due to the small mass of this intermediate that accumulates in the reaction mixture. These LC/MS data and the chromatographic and spectral properties of the epidermal ¹⁴C products as shown in Table 1 are in agreement and confirm the proposed structural identities of products I to IV. These are: 12-oxo-ETrE, 12-HETrE, 12-oxo-ETE, and 12-HETE, respectively.

Evidence That Lipoxygenases in Mouse Epidermal Microsomes, Rather Than Cytochromes P450, Have Major Roles in 12-HETrE Biosynthesis. To investigate precursor-product relationships, we measured the accumulation of epidermal ¹⁴C products I to IV over time in the presence of NADPH (Fig. 5). 12-Hydroxyeicosatetraenoic acid (IV) was by far the most abundant product. The mass of 12-HETE increased continuously for 40 min, indicating 12-HETE is the major end product of this reaction (Fig. 5A). The second most abundant product measured was 12-hydroperoxy eicosatetraenoic acid (12-HpETE), the immediate product of 12-lipoxygenase. However, the mass of 12-HpETE decreased after 10 min to very low levels. The mass of 12-HETrE (II) increased continuously over 40 min, indicating 12-HETrE is also an end product (Fig. 5A and expanded scale in Fig. 5B). The initial increase in 12-oxo-ETE (III) and 12-oxo-ETrE (I) accumulation slowed markedly or began to decrease after 10 min, resembling the pattern observed for 12-HpETE, consistent with their roles as metabolic intermediates.

Two end products were formed from arachidonate, 12-HETE and lesser amounts of 12-HETrE (Fig. 5). To ascertain whether P450 proteins were involved in their synthesis, we used 1-aminobenzotriazole, a mechanism-based inhibitor of cytochrome P450 having low isoform selectivity. Pretreatment of mouse epidermal microsomes with 100 to 1000 μM 1-aminobenzotriazole did not selectively block NADPH-dependent product formation (data not shown), nor could we detect a measurable decrease in 12-HETE accumulation. We conclude that epidermal lipoxygenases, rather than P450 proteins, likely have prominent roles in 12-HETrE synthesis, along with other unidentified enzymes responsible for the NADPH-dependent conversion of 12-oxo-ETE to 12-oxo-ETrE and 12-HETrE.

Discussion

Regioisomeric EETs generated from endogenous arachidonic acid are present in mouse epidermal tissue and differentiating epidermal cell cultures, indicating the presence of catalytically active epoxygenases (Keeney et al., 1998; Du et al., 2005). Because existing evidence suggests mouse CYP2B19 is mainly responsible for cellular EET formation in mouse skin (Du et al., 2005), we set out to demonstrate that this keratinocyte-specific epoxygenase in epidermal microsomes is responsible for generating EETs from [1-¹⁴C]arachidonate in vitro. Instead of EET formation, we identified a NADPH-dependent pathway leading to biosynthesis of 12-HETrE. This is the first report showing this pathway is active in mouse skin.

A novel feature of mouse epidermal microsomes is that we were able to identify and characterize all four intermediates in synthesis of 12-HETrE from arachidonate: 12-HETE/12-HpETE, 12-oxo-ETE, and 12-oxo-ETrE. Second, we found no evidence for formation of 15-hydroxyeicosatrienoic acid isomers, unlike rat epidermal microsomes, which generate 15-hydroxy-5,8,11-eicosatrienoic acid from arachidonate (Van Wauwe et al., 1991, 1992). This might represent a species difference, or the levels of this regioisomer were just too low to be detected in our assay system. In human skin, HETrE formation has been reported from dietary polyunsaturated fatty acid but not from arachidonic acid. For example, human epidermal cells can generate 15-hydroxy-8,11,13-eicosatrienoic acid from γ-linolenic acid (Ziboh et al., 2000).

Even though EET biosynthesis was not detected in vitro, CYP2B19 is expressed in mouse epidermal keratinocytes; CYP2B19-derived peptides were identified unequivocally by mass spectrometry, following immunoaffinity chromatography (Du et al., 2005). In addition to explanations already discussed about why EET biosynthesis was not detected, we question whether epidermal epoxygenases utilize exogenous arachidonic acid efficiently in the microsomal environment. Cyclooxygenase (PTGS1 and PTGS2) activity levels differed depending on whether substrate was exogenous or endogenous arachidonic acid (Chulada et al., 1996). In our system, exogenous arachidonate was utilized much more efficiently by epidermal lipoxygenases and cyclooxygenases than epoxygenases. Even though soluble and microsomal epoxide hydrolases are expressed in the epidermis (Winder et al., 1993), we found little evidence for hydrolysis of EETs to vicinal diols using an inhibitor of soluble epoxide hydrolase, the form mainly responsible for hydrolysis of epoxy fatty acids (Newman et al., 2005).

The NADPH-dependent products I and II (12-oxo-ETrE and 12-HETrE, respectively) appeared to form at the expense of product III (12-oxo-ETE). This was the first evidence for a precursor-product relationship. The time-dependent accumulation and disappearance of epidermal ¹⁴C products (I–IV) lend credibility to the idea that HETrE biosynthesis in mouse epidermal microsomes shares some of the same intermediates as reported for bovine corneal and rat epidermal microsomes (Van Wauwe et al., 1992; Yamamoto et al., 1994). In the mouse epidermal microsome system, we identified 12-HETE and 12-HETrE as end products and 12-oxo-ETE and 12-oxo-ETrE as intermediates in the formation of 12-HETrE.

Mechanistically, 12-HETrE synthesis could involve lipoxygenases and/or cytochromes P450 (Nishimura et al., 1991; Van Wauwe et al., 1991, 1992; Yamamoto et al., 1994). CYP4B1 has been implicated in 12-HETrE formation in the cornea (Mastyugin et al., 1999). We were unable to establish specific evidence implicating a P450 protein in 12-HETrE formation, but neither could we specifically rule out this possibility.

Epidermal 12-lipoxygenase activities were likely involved in 12-oxo-ETE biosynthesis since highest levels of this intermediate were observed in the absence of NADPH. Of six functional lipoxygenase genes in humans and seven in mice, several, including platelet-type and epidermis-type 12S-lipoxygenase and 12R-lipoxygenase, are expressed in skin, potentially contributing to 12-oxo-ETE formation (de Laclos et al., 1987; Antón and Vila, 2000; Funk et al., 2002; Yoshimoto and Takahashi, 2002). That 12-HETE was the most abundant end product is consistent with previous studies establishing 12-lipoxygenase as the major lipoxygenase activity in mouse skin microsomes (Nakadate et al., 1986).

Enzymes responsible for the NADPH-dependent conversion of 12-oxo-ETE to 12-oxo-ETrE and 12-HETrE have not been studied in the epidermis. However, in porcine leukocytes, a microsomal NAD⁺-dependent dehydrogenase activity was characterized that generates 12-oxo-ETE from 12-HETE. This oxidation step was prerequisite for maximal activity of a cytosolic NADH-dependent 10,11-reductase (Wainwright et al., 1990; Wainwright and Powell, 1991). Involvement of a cytosolic reductase in mouse epidermis is contradicted by the observation that microsomes prepared from epidermis and full-thickness skin (epidermis and dermis) were competent to generate 12-HETrE; however, we cannot rule out minor cytosolic contamination. The 12-/15-oxo-ETE and 12-/15-oxo-ETrE intermediates are common to diverse cell systems, suggesting that pathways for HETrE are conserved in mammalian skin, cornea, leukocytes, and quite possibly many other tissues. Important differences might be found in the mechanisms regulating production of regioisomeric HETrEs or biologically active intermediates.

In differentiating keratinocytes, P450-derived EETs are implicated in mechanisms regulating transglutaminase enzyme activities, which are critical for normal epidermal cornification (Ladd et al., 2003). Is epidermal HETrE biosynthesis physiological relevant? In the cornea, 12R-HETrE formation is enhanced by injury and hypoxia, and its proinflammatory properties include vasodilation of blood vessels, neutrophil chemoattraction, endothelial cell proliferation, and neovascularization (Murphy et al., 1988; Stoltz and Schwartzman, 1997). Both 12-HETE enantiomers are substrate for 12-HETrE biosynthesis, but only 12R-HETrE has significant proinflammatory properties (Yamamoto et al., 1994; Stoltz and Schwartzman, 1997). In guinea pig skin, 12R-HETrE is also proinflammatory, enhancing delayed-type hypersensitivity inflammatory reactions at dosages as low as 1 fmol (Conners et al., 1995). Neither of the 12-HETE enantiomers nor 12S-HETrE had this activity.

The extent that epidermally derived 12-HETE serves as precursor for 12-HETrE biosynthesis in human skin is unknown since this pathway has not been studied in human epidermal cells, nor have cellular 12-HETrE levels been reported for normal or diseased skin. In hyperproliferative dermatoses (e.g., psoriatic), 12-HETE accumulates to greater levels in lesional (versus normal) skin (Hammarström et al., 1979). The 12R-HETE enantiomer predominates and is more chemotactic (Woollard, 1986; for review, see Ikai, 1999; Fogh and Kragballe, 2000).

Results of recent studies suggest the proinflammatory activities of 12-HETrE are counterbalanced by the anti-inflammatory and antiproliferative activities of 15-HETrE (Ziboh et al., 2000) and that these lipid mediators may utilize different signaling mechanisms. High-affinity binding sites characterized in microvessel endothelial cells potentially mediate the proinflammatory actions of 12R-HETrE (Stoltz and Schwartzman, 1997). The 15-HETrE regioisomer was found incorporated into phospholipids, and when released by phospholipase C as 15-HETrE-diacylglycerol, this lipid moiety modulated protein kinase C activities (for review, see Ziboh et al., 2000).

Important future studies are to establish whether 12-/15-HETrE biosynthetic pathways are operative in human epidermal cells, to characterize the regio- and stereochemistry of the eicosanoids formed and their biological activities in human skin. Opposing pro- and anti-inflammatory activities for 12- and 15-HETrE make sense biologically to regulate cutaneous inflammatory and immune responses. In this regard, it was proposed that 15-HETE might influence the development of psoriatic lesions by its ability to inhibit synthesis of the proinflammatory lipids 5- and 12-HETE (for review, see Ikai, 1999). Again, it is unclear to what extent lipoxygenase-derived products are utilized as precursor for hydroxyeicosatrienoic acid biosynthesis or to what extent this secondary metabolism contributes to observed biological effects. In addition to modulating cutaneous inflammatory responses, it seems likely that regioisomeric HETrEs and biosynthetic intermediates have additional activities that remain to be discovered.