Biogenic Synthesis, Purification, and Chemical Characterization of Anti-inflammatory Resolvins Derived from Docosapentaenoic Acid (DPAn-6)

Abstract

Enzymatically oxygenated derivatives of the ω-3 fatty acids cis-4,7,10,13,16,19-docosahexaenoic acid (DHA) and cis-5,8,11,14,17-eicosapentaenoic acid, known as resolvins, have potent inflammation resolution activity (Serhan, C. N., Clish, C. B., Brannon, J., Colgan, S. P., Chiang, N., and Gronert, K. (2000) J. Exp. Med. 192, 1197–1204; Hong, S., Gronert, K., Devchand, P. R., Moussignac, R., and Serhan, C. N. (2003) J. Biol. Chem. 278, 14677–14687). Our objective was to determine whether similar derivatives are enzymatically synthesized from other C-22 fatty acids and whether these molecules possess inflammation resolution properties. The reaction of DHA, DPAn-3, and DPAn-6 with 5-, 12-, and 15-lipoxygenases produced oxylipins, which were identified and characterized by liquid chromatography coupled with tandem mass-spectrometry. DPAn-6 and DPAn-3 proved to be good substrates for 15-lipoxygenase. 15-Lipoxygenase proved to be the most efficient enzyme of the three tested for conversion of long chain polyunsaturated fatty acids to corresponding oxylipins. Since DPAn-6 is a major component of Martek DHA-S™ oil, we focused our attention on reaction products obtained from the DPAn-6 and 15-lipoxygenase reaction. (17S)-hydroxy-DPAn-6 and (10,17S)-dihydroxy-DPAn-6 were the main products of this reaction. These compounds were purified by preparatory high performance liquid chromatography techniques and further characterized by NMR, UV spectrophotometry, and tandem mass spectrometry. We tested both compounds in two animal models of acute inflammation and demonstrated that both compounds are potent anti-inflammatory agents that are active on local intravenous as well as oral administration. These oxygenated DPAn-6 compounds can thus be categorized as a new class of DPAn-6-derived resolvins.

Enzymatically formed oxygenation products of C-20 and C-22 long chain polyunsaturated fatty acids (LC-PUFAs),4 have important biological roles in inflammation, allergies, and blood clotting and are thus believed to have therapeutic potential in several chronic immune diseases (1–10) Several biologically important products of cis-5,8,11,14-eicosatetraenoic acid/arachidonic acid (ARA), cis-5,8,11,14,17-eicosapentaenoic acid (EPA), and cis-4,7,10,13,16,19-docosahexaenoic acid (DHA) have been described (4, 11, 12). Proinflammatory oxylipins, such as leukotrienes and some prostaglandins, are derived from ARA, an ω-6 fatty acid. Interestingly, the same fatty acid also serves as a precursor to anti-inflammatory or proresolution molecules like lipoxins (13, 14). Stable analogues of lipoxins are being developed as drugs for asthma and other inflammatory airway diseases (15, 16). Oxylipins derived from ω-3 fatty acids, such as DHA and EPA, known as resolvins, are primarily anti-inflammatory in nature (17). EPA acts as a precursor to the E-series resolvins that have shown potential in the treatment of colitis, arthritis, and periodontitis (18–20). The resolvins of the D-series derived from DHA are useful as neuroprotective agents. 10,17-Dihydroxy-4,7,11,13,15,19-docosahexaenoic acid (10,17-HDHA) or neuroprotectin D1 is a resolvin that is formed endogenously in the human brain and eye and is believed to exert its protective effect against cell injury-induced oxidative stress (21–23).

The main enzymes responsible for the production of these oxygenated LC-PUFA products are primarily lipoxygenases and, in addition, cyclo-oxygenases and cytochromes P450. These enzymes produce oxylipins via transcellular activity, often involving multiple cell types (24). This activity mainly results in mono-, di-, and tri-hydroxylation products of fatty acids that have varying potencies, depending on the exact structure of the compound. Lipoxygenases are non-heme, iron-containing dioxygenases that catalyze the regioselective and enantioselective oxidation of polyunsaturated fatty acids containing one or more cis,cis-1,4-pentadienoic moieties to give the corresponding hydroperoxy derivatives (25, 26). We thus considered that, in addition to DHA and EPA, other C-22 PUFAs containing such methylene interrupted double bonds may also be substrates for lipoxygenases and that resulting products may have anti-inflammatory activity similar to DHA-derived resolvins. DPAn-6 (cis-4,7,10,13,16-docosapentaenoic acid) is present in algal oils, and recent studies have demonstrated that this fatty acid has anti-inflammatory activities in vitro and, in conjunction with DHA, also has anti-inflammatory activity in vivo.5 Also, it has been suggested that a combination of DHA and DPAn-6 could be a beneficial natural therapy in neuroinflammatory conditions like Alzheimer disease. Specifically, in a 3×Tg-AD mouse model of Alzheimer disease, DPAn-6 was shown to reduce levels of early stage phospho-Tau epitopes, which in turn correlated with a reduction in phosphorylated c-Jun N-terminal kinase, a putative Tau kinase (27). Although the precise mechanism of action of DPAn-6 in these inflammatory milieus is not known, it suggests a possible role for oxylipin products of DPAn-6 in resolution of inflammation. Also, another LC-PUFA, DPAn-3 (cis-7,10,13,16,19-docosapentaenoic acid) usually present along with DHA and EPA in marine oils is known to be a potent inhibitor of platelet aggregation (28–30). In addition, this LC-PUFA has a potent inhibitory effect on angiogenesis through the suppression of VEGFR-2 (vascular endothelial-cell growth factor receptor 2) expression. Angiogenesis is known to contribute to tumor growth, inflammation, and microangiopathy, again pointing to the possibility that anti-inflammatory activity of DPAn-3 might be mediated through resolvin-like products as in the case of DHA and EPA (31).

The purpose of this research was to determine whether oxylipins are formed from the C-22 LC-PUFAs, DPAn-6 and DPAn-3, by lipoxygenase activity; to compare them to products formed from DHA; to chemically characterize products; to purify key oxylipin products from the DPAn-6/15-lipoxygenase reaction; and to test whether these compounds have resolvin-like anti-inflammatory activity. This research also sets the stage for preparation and isolation of a wide range of other C-22 oxylipins that could be evaluated as potential anti-inflammatory compounds.

EXPERIMENTAL PROCEDURES

Materials—5-lipoxygenase from potato, 12-lipoxygenase from porcine leukocytes, and (17S)-hydroxy-4,7,10,13,15,19-docosahexaenoic acid ((17S)-HDHA) were from Cayman Chemicals (Ann Arbor, MI). Tumor necrosis factor-α (TNFα) was from Peprotech, Inc. (Rocky Hill, NJ). All fatty acids were from Nu-Chek Prep (Elysian, MN). Filtration units were from Nalgene (Rochester, NY). Acrodisc syringe filters were from Pall Life Sciences (Ann Arbor, MI). All Luna high performance liquid chromatography (HPLC) columns were from Phenomenex (Torrance CA), Chiralpak-IA was from Chiral Technologies (West Chester, PA), Chirobiotic-T was from Astec (Whippany, NJ), and all HPLC solvents were from Fisher. All other reagents and materials were from Sigma, whereas animals were from Harlan (Indianapolis, IN).

Reaction of DHA, DPAn-3, and DPAn-6 with 15-Lipoxygenase for Comparison of Efficiency of Conversion—Soybean 15-lipoxygenase (Type 1-B) at a final concentration of 524 units/ml was mixed into 100 μm solutions of DHA, DPAn-6, or DPAn-3 fatty acids in 0.05 m sodium borate buffer, pH 9.0, and the reaction mixtures were incubated at 0 °C. The appearance of the monohydroxy-conjugated diene derivatives of the fatty acids was monitored through absorbance at 237 nm. Conjugated diene products were quantified by UV spectrophotometry using an extinction coefficient of 28,000/m·cm (32).

Enzyme Kinetics Measurements of 15-Lipoxygenase with DHA, DPAn-3, and DPAn-6—Type 1-B soybean lipoxygenase (39.8 units/ml) was used to dioxygenate substrates (DPAn-6, DPAn-3, and DHA at 5–100 μm) at 22 °C in oxygen-saturated 0.05 m sodium borate buffer (pH 9.0) with 4.52 × 10^–3% (v/v) Tween 20. Final oxygen concentrations exceeded 750 mm, measured using a Clark type oxygen electrode. The probe was calibrated by bubbling air and directly measuring the concentration with immobilized fluorescent ruthenium (OxySense 200T; OxySense, Dallas, TX). Approximately 38 s after the enzyme addition, the hydroperoxide product formation was monitored via UV absorbance at 236 nm (as described above) at 15-s intervals for 600 s (DU800 spectrophotometer; Beckman Coulter, Fullerton, CA). The read average time was 0.2 s. All experiments were repeated in triplicate. Product formation curves were smoothed with a negative exponential transform sampling all of the data with a power of 10 calculating intervals every second. Product formation curves showed an initial lag period followed by a maximum in rate (r_max). Although this is not the classical V_0max, previous authors have shown that it satisfies the steady-state assumption for application of the Michaelis-Menten kinetic model (33, 34). This model, r = V_max[S]/(K_m + [S]), was used to elucidate the kinetic parameters. SigmaPlot 11.0 with the Enzyme Kinetic Module 1.3 (Systat Software, Inc., San Jose, CA) was used to perform the least squares fit of the aforementioned model.

Reaction of DHA, DPAn-3, and DPAn-6 with 5-, 12-, and 15-Lipoxygenase for Identification of Reaction Products—For the 5-lipoxygenase reaction, 200 μl of 10 units/μl potato 5-lipoxygenase was added to a 10-ml reaction mixture containing 100 μm fatty acid in 0.05 m Na-MES buffer, pH 6.3, 100 μm SDS, and 0.02% decaethylene glycol monododecyl ether. For the 12-lipoxygenase reaction, 100 μl of 0.75 units/μl porcine leukocyte-derived 12-lipoxygenase was added to a 10-ml solution containing 100 μm fatty acid in 0.1 m Tris-HCl, pH 7.5, 5 mm EDTA, and 0.03% Tween 20. For the 15-lipoxygenase reaction, 100 μl of 131 kilounits/ml soybean 15-lipoxygenase (Type 1-B) was added to a 10-ml reaction mixture containing 100 μm fatty acids in 0.05 m sodium borate buffer, pH 9.0. Control reactions were run for all three enzymes under identical conditions without adding any enzyme. All reaction mixtures were stirred for 30 min at 4 °C. Reaction products were reduced with sodium borohydride and acidified with glacial acetic before extraction on solid phase DSC-18 cartridges. Final elutions were done using anhydrous ethanol and reaction products analyzed by LC-MS/MS (liquid chromatography coupled with tandem mass spectrometry). The reaction of 15-lipoxygenase with DPAn-6 described above was scaled-up for the production of 17-hydroxy-4,7,10,13,15-docosapentaenoic acid (17-HDPAn-6) and 10,17-dihydroxy-4,7,11,13,15-docosapentaenoic acid (10,17-HDPAn-6) from 1 g of DPAn-6.

Purification of 17-HDPAn-6 and 10,17-HDPAn-6—All purifications were carried out at a flow rate of 47 ml/min at room temperature on a Luna C18 (2), 10 μm, 100 Å, 50 × 250-mm column, set up on an Agilent 1200 preparatory HPLC unit, equipped with a diode array detector. The purification method involved elution with a mixture of solvents A and B. Solvent A consisted of 70% water, 30% methanol, and 0.2% ammonium acetate, and solvent B consisted of 100% acetonitrile. The method used for elution of compounds was 48% B at 0 min, 48% B at 25 min, 90% B at 35 min, 90% B at 50 min, 48% B at 52 min, and 48% B at 65 min. Chromatograms were monitored simultaneously at 236 and 270 nm, and fractions corresponding to peaks of interest were collected and pooled from several runs. Purity of fractions was monitored by UV spectrophotometery as well as by LC-MS. Fractions were concentrated using a rotary evaporator, and compounds were extracted with ether. Ether extracts were concentrated and treated with anhydrous sodium sulfate to remove any traces of water. This solution was then filtered using Acrodisc syringe filters (Bulk Acrodisc CR 13 mm with 0.2 μm polytetrafluoroethylene membrane, HPLC-certified). Purity was ascertained by LC-MS and LC-MS/MS.

UV Spectrophotometry—UV spectrophotometry was carried out on either an Agilent 8453 (Agilent Technologies, Santa Clara, CA) or a Cary 4000 UV-visible spectrophotometer (Varian Inc., Palo Alto, CA) or a DU800 spectrophotometer (Beckman Coulter). Extinction coefficients of 17-HDPAn-6 and 10,17-diHDPAn-6 were determined in ethanol.

LC-MS and LC-MS/MS—All routine reaction and purity analysis was done on a Hewlett Packard model 1100 liquid chromatography instrument interfaced with an electrospray ionization mass elective detector. Tandem MS/MS analysis of the products of the reactions of fatty acids and different lipoxygenases was performed on an Agilent 1100 Series HPLC instrument (San Paulo, CA) interfaced with an Esquire 3000 ion trap mass spectrometer equipped with an electrospray ionization source (Bruker Daltonics, Billerica, MA) or on an Agilent 1200 Series HPLC instrument interfaced with an Agilent LC/MSD Trap XCT Plus (San Paulo, CA). Mass spectrometers were operated in negative ion detection mode. Nitrogen was used as nebulizing and drying gas with nebulizer pressure at 20 p.s.i. and drying gas flow rate of 7 liters/min. The interface temperature was maintained at 330 °C. The chromatographic separation was carried out on a LUNA C18(2) column (250 × 4.6 mm, 5 μm, 100 Å) using the same mobile phase described for purification of the compounds under similar conditions at a flow rate of 0.4 ml/min.

NMR Experiments—All NMR experiments were conducted by NMRServices (Rochester, NY). 25–50 mg of 17-HDPAn-6 or 10,17-HDPAn-6 was dissolved in CDCl₃ and used for NMR data acquisition. All NMR data were acquired at 298 K on a Bruker model Avance spectrometer operating at a 500-MHz proton NMR frequency. One-dimensional ¹H, one-dimensional ¹³C(¹H-decoupled), two-dimensional ¹H-¹H (correlation spectroscopy), two-dimensional ¹H-¹H DQFCOSY (double-quantum-filtered correlation spectroscopy), two-dimensional ¹H-¹³C-HSQC (heteronuclear single quantum coherence), and two-dimensional ¹H-¹³C-HMBC (heterononuclear multiple bond-correlation) NMR data were acquired for 17 HDPAn-6. One-dimensional ¹H, one-dimensional ¹³C (¹H-decoupled), and two-dimensional ¹H-¹H COSY were acquired for 10,17-HDPAn-6. All data were processed using Bruker software, and NMR spectra predictions were done using ChemDraw Ultra 10.0 and ACDlabs software.

Chirality—17-HDHA was prepared the same way as 17-HDPAn-6 except that DHA was used as a substrate. Either 17-HDHA or 17-HDPAn-6 was converted to 17-oxo-derivatives by dissolving 15 mg of each compound in anhydrous dichloromethane and then oxidizing the compounds at 0 °C with 15 mg of Dess-Martin's reagent. The 17-oxo-derivatives were then reduced using sodium borohydride to give the corresponding racemic mixtures. The racemic mixtures were then separated on a Chiralpak-IA column using methanol/water (80:20) as the mobile phase at a flow rate of 0.6 ml/min at 25 °C. The peaks were assigned as R or S by using (17S)-HDHA as a reference standard. The same protocol was followed for 17-HDPAn-6, and the order of elution of the two isomers was assumed to be the same as for 17-HDHA isomers. The Chirobiotic-T column was tested with supplier-recommended mobile phases. Vibrational circular dichroism (VCD) studies were carried out by Biotools Inc. (Syracuse, NY) for 17-HDPAn-6.

Log D_{pH 7.4} and Solubility—Log D_{pH 7.4} and solubility studies were conducted by Cerep (Seattle, WA). All predictions of various physicochemical parameters were made using ACDLabs software. Log P was calculated based on the formula, log P = log D + (pH – pK_a) (35) Aqueous solubility (μm) was determined by comparing the peak area of the principal peak in a calibration standard (200 μm) containing organic solvent (methanol/water, 60/40, v/v) with the peak area of the corresponding peak in a buffer sample. Reference compounds used for comparing solubility were diethylstilbestrol, haloperidol, ketoconazole, metoprolol tartrate, phenytoin, rifampicin, simvastatin, and tamoxifen. Log D_{pH 7.4} was determined by the shake flask method using octanol and PBS, pH 7.4 (35). The total amount of compound was determined as the peak area of the principal peak in a calibration standard (100 μm) containing organic solvent (methanol/water, 60/40, v/v). The amount of compound in the buffer was determined as the combined, volume-corrected, and weighted areas of the corresponding peaks in the aqueous phases of three organic aqueous samples of different composition. An automated weighting system was used to ensure the preferred use of raw data from those samples with well quantifiable peak signals. The amount of the compound in the organic phase was calculated by subtraction. Subsequently, log D was calculated as log₁₀ of the amount of compound in the organic phase divided by the amount of compound in the aqueous phase.

Chemical Stability Studies—17-HDPAn-6 and 10,17-HDPAn-6 were incubated at a concentration of 10 μm in both PBS and ethanol at 4 °C for 14 days and at 22 °C for 24 h and at 37 °C for 2 h, and peak heights were monitored by LC-MS/MS. For the 14-day stability studies, samples were analyzed at 0, 2, 4, 7, 9, 11, and 14 days. Samples were analyzed at 0 and 24 h for the 22 °C study and at 0 and 2 h for the 37 °C study.

Animal Studies—All animal studies were performed in accordance with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the National Institutes of Health and/or the Final Rules of the Animal Welfare Act Regulation (9CFR).

Mouse Air Pouch Study—To generate the dorsal air pouches, 6–8-week-old female C57Bl/6N mice were anesthetized, and 6 ml of sterile air was injected subcutaneously in the back (36). Pouches were reinflated with 3 ml of sterile air 3 days later. Six days after initial pouch formation, 100 ng of the DPAn-6 oxylipin test compounds or (17S)-HDHA were administered to 10 animals per group in 1 ml of PBS, pH 7.4 (∼0.005 mg/kg) by intrapouch injection. Within 5 min of compound administration, 100 ng of TNFα in 0.1 ml of PBS was injected into the pouches to initiate an inflammatory response. Indomethacin was administered intraperitoneally at 2 mg/kg 30 min prior to the TNFα challenge as a positive control. Four h after the TNFα challenge, pouch exudates were harvested, and leukocytes in the exudates were counted. A sampling of exudate cells from each animal was centrifuged onto microscope slides, and the cells were airdried and Geimsa-stained to quantify granulocyte and mononuclear leukocytes.

Rat Hind Paw Edema Studies—Hind paw edema studies were conducted on male Sprague-Dawley rats (170–200 g, 6–8 animals/group) using cross-linked λ carrageenan as the inflammatory stimulus (37). Researchers were masked to the treatment groups in all hind paw studies. The first experiment involved administering DPAn-6-derived resolvins either intravenously (5 μg/animal, ∼0.025 mg/kg) in the tail vein in 0.1 ml of PBS 2 min prior to carrageenan challenge or per os (by oral gavage) in up to 4 ml of PBS (25 μg/animal, ∼0.125 mg/kg) 30 min prior to carrageenan challenge. Indomethacin (5 mg/kg administered intraperitoneally 30 min prior to challenge) and biogenically synthesized (17S)-HDHA (5 μg/animal administered intravenously 2 min prior to challenge) served as positive and DHA resolvin controls, respectively, in this experiment. Rats were injected with 0.1 ml of 10 mg/ml carrageenan solution (prepared in water at least 2 weeks in advance of the experiment) in the subplantar region of the right hind paw. Paw volumes were measured by water volume displacement at 0, 2, 4, 6, and 24 h post-carrageenan challenge.

The second experiment involved evaluating oral dose-response behavior of the two DPAn-6 resolvins. All compounds, including the controls, were administered by oral gavage (4 ml/animal) 30 min prior to the carrageenan challenge. The two DPAn-6 oxylipins were diluted in PBS, pH 7.4, and dosed at 2.5, 25, and 250 μg/animal (∼0.0125, 0.125, and 1.25 mg/kg body weight), and the biogenically synthesized resolvin control 17-HDHA was given at a single dose of 25 μg/animal (0.125 mg/kg). Indomethacin was administered at 6 mg/animal (30 mg/kg). Carrageenan was administered as in the first rat hind paw experiment. Paw edema volume was measured 2, 4, and 6 h after the carrageenan challenge. Paw edema volume was calculated by subtracting right paw volume just prior to carrageenan injection from the post-carrageenan-treated volume of the same paw.

Statistical Analyses—Statistical analyses for animal studies were conducted in GraphPad Prism version 4 (San Diego, CA) using Student's t tests or ANOVA with Dunnett's post test to compare treatment groups with control. Data are presented as means ± S.D.

RESULTS

Substrate Comparison with 15-Lipoxygenase—Fig. 1 shows percentage conversion of DHA, DPAn-3, and DPAn-6 upon treatment with soybean lipoxygenase at 0 °C. As can be seen, under these conditions, 100% of DPAn-6 was efficiently converted to its conjugated diene derivative, whereas about 85% of DPAn-3 and 50% of DHA were converted to their respective conjugated diene derivatives by 15-lipoxygenase. These results were also corroborated by LC/MS peak analysis, which showed that under these conditions, a single product with a conjugated diene was formed, and this corresponded to the 17-hydroperoxy derivative in all three cases. In addition, we also assessed enzyme kinetics parameters of the three substrates with 15-lipoxygenase at 22 °C. Fig. 2 shows the enzyme kinetics data with curve fits, and Table 1 shows the kinetic parameters. Over the substrate concentrations analyzed, the kinetic behavior was Michaelis-Menten. Although DPAn-3 exhibited the highest value for V_max, the efficiency with which the enzyme converted DPAn-6 to products was higher than that for DHA and DPAn-3.

FIGURE 1.

Comparison of DHA, DPAn-3, and DPAn-6 as substrates for soybean lipoxygenase. Fatty acids (DHA (○), DPAn-3 (□), and DPAn-6 (▴)) at a concentration of 100 μm were incubated with enzyme at 0 °C, as described under “Experimental Procedures.” Absorbance at 237 nm was used to estimate the amount of product formed, and theoretical 100% conversion was calculated based on an extinction coefficient of 28,000/m·cm in all three cases. The percentage conversions were verified by LC/MS. The graph shows results from replicate measurements. The error bars represent ranges.

FIGURE 2.

Enzyme kinetics measurements and curve fits for DPAn-6, DPAn-3, and DHA. The reaction was carried out at 22 °C with 39.8 units/ml of the enzyme, 4.52 × 10^–3 % (v/v) Tween 20 in oxygen-saturated 0.05 m sodium borate buffer (pH 9.0), and products were quantified by measuring absorbance at 236 nm. Shown are actual (circles) and predicted (trace) hydroperoxide formation rates as a function of substrate concentration for DPAn-3 (A), DPAn-6 (B), and DHA (C). Actual results are averages of triplicates, presented with S.E. bars.

TABLE 1.

Kinetic parameters of PUFA substrates with 15-lipoxygenase

Substrates (5–100 μm) were incubated with 39.8 units/ml of 15-lipoxygenase, 4.52 × 10^-3% (v/v) Tween 20 in oxygen-saturated 0.05 m sodium borate buffer (pH 9.0), and products were quantified using absorbance at 236 nm. Curve fits were created using the Michaelis-Menten model.

Parameter	DPAn-3	DPAn-6	DHA
Vmax (μm/min)	672.9 ± 31.8	487.1 ± 13.9	481.9 ± 21.4
Km (μm)	36.2 ± 4.1	14.3 ± 1.5	20.4 ± 2.8
Vmax/Km (min-1)	18.6 ± 2.3	34.1 ± 3.7	23.6 ± 3.4
R2	0.93	0.90	0.85

Products Formed upon Reaction of Fatty Acids with 5-, 12-, and 15-Lipoxygenase—Tables 2, 3, 4 give a summary of products formed when DHA, DPAn-3, and DPAn-6 are reacted with the various lipoxygenases. In all cases, the enzymes produced hydroperoxy compounds, which were then reduced nonenzymatically by sodium borohydride to the corresponding hydroxy compounds. The hydroperoxy derivatives (data not shown) were detected in nonreduced reactions by LC-MS/MS techniques. Final reduced products were identified based on parent ion masses and diagnostic fragments. MS spectra of all monohydroxy compounds showed [M-H], [M-H]-H₂O, [M-H]-CO₂, and [M-H]-CO₂,-H₂O, which were 345, 327, 301, and 283 Da, respectively, for monohydroxy-DPAn-6 and -DPAn-3 derivatives and were 343, 325, 299, and 281 Da for monohydroxy-DHA derivatives. MS spectra of all dihydroxy compounds showed [M-H], [M-H]-H₂O, [M-H]-CO₂, [M-H]-CO₂,-H₂O, [M-H]-2H₂O, and [M-H]-CO₂,-2H₂O, which were 361, 343, 317, 325, and 281 Da, respectively, for dihydroxy-DPAn-3 and DPAn-6 derivatives and were 359, 341, 315, 323, and 279 for dihydroxy-DHA derivatives. All diagnostic fragments that help in ascertaining positions of oxygenation are outlined in Tables 2, 3, 4.

TABLE 2.

Products formed upon incubation of DHA, DPAn-3, and DPAn-6 with soybean 15-lipoxygenase

DHA, DPAn-3, and DPAn-6 were incubated with soybean 15-lipoxygenase in 0.05 m sodium borate buffer, pH 9.0, at a concentration of 100 μm, at 4 °C for 30 min, and products were reduced to the corresponding hydroxy compounds using sodium borohydride. Reactions were analyzed and product structures were elucidated based on LC-MS/MS analysis of diagnostic fragments. Diagnostic fragments that helped in ascertaining positions of hydroxylation are shown in parentheses.

TABLE 3.

Products formed upon incubation of DHA, DPAn-3, and DPAn-6 with porcine 12-lipoxygenase

DHA, DPAn-3, and DPAn-6 were incubated with 12-lipoxygenase obtained from porcine leukocytes in 0.1 m Tris-HCl, pH 7.5, 5 mm EDTA, and 0.03% Tween 20, at a concentration of 100 μm, at 4 °C for 30 min, and products were reduced to the corresponding hydroxy compounds using sodium borohydride. Reactions were analyzed by LC-MS/MS, and product structures were elucidated based on analysis of diagnostic fragments. Diagnostic fragments that helped in ascertaining positions of hydroxylation are shown in parentheses.

TABLE 4.

Products formed upon incubation of DHA, DPAn-3, and DPAn-6 with potato 5-lipoxygenase

DHA, DPAn-3, and DPAn-6 were incubated with potato 5-lipoxygenase in 0.05 m Na-MES buffer, pH 6.3, 0.02% C12E10 at a concentration of 100 μm, at 4 °C for 30 min, and products reduced to the corresponding hydroxy compounds using sodium borohydride. Reactions were analyzed by LC-MS/MS, and product structures were elucidated based on analysis of diagnostic fragments. Diagnostic fragments that helped in ascertaining positions of hydroxylation are shown in parentheses.

Of all of the enzymes tested, 15-lipoxygenase was the most efficient in converting fatty acids to oxylipins, whereas 5-lipoxygenase and 12-lipoxygenase converted less than 10% of the fatty acids to oxylipin-like compounds. Overall, many of the products characterized upon using DPAn-6 or DPAn-3 as substrates were similar to those obtained from DHA, as reported earlier (1, 3, 5). As expected, the 15-lipoxygenase products were primarily oxygenated at the C-17-position, ultimately producing 17-HDHA, 17-HDPAn-6, and 17-HDPAn-3 from the respective fatty acids. These compounds had a conjugated diene structure, with an absorbance maximum between 234 and 238 nm. In addition, 10,17-dihydroxy and 7,17-dihydroxy derivatives were produced from all three fatty acids. The 10,17-dihydroxy derivatives had an absorbance maximum at 270 nm with characteristic shoulders at 260 and 280 nm (±2 nm), which are characteristic of a conjugated triene structure. The 7,17-dihydroxy derivatives had an absorbance maximum at ∼242 ± 2 nm, characteristic of two conjugated dienes interrupted by a methylene group. The main product (>90%) formed upon reacting 12-lipoxygenase with fatty acids was the 14-hydroxy derivative (hydroxylation at the ω-9-position) in all three cases. In addition, several other mono as well as dihydroxy derivatives were detected (see Table 3). When 5-lipoxygenase was used, the main product (>90%) formed was due to oxygenation at the ω-13-position, resulting in the 10-hydroxy derivative. Other products were also seen (see Table 4). The 10,20-dihydroxy derivative was formed only from DHA and DPAn-3 and not from DPAn-6, since DPAn-6 lacks a double bond at C-19. None of the controls for any of the enzymatic reactions showed any appreciable levels of oxylipins, suggesting that there was no appreciable amount of nonenzymatic autoxidation under the conditions tested.

Gram Scale Synthesis and Purification of 15-Lipoxygenase DPAn-6 Resolvins—Fig. 3 shows the LC profiles of products obtained when 1 g of DPAn-6 was reacted with 15-lipoxygenase. The main products that were formed were 17-HDPAn-6 (∼85%), 10,17-HDPAn-6 (∼4%), and 7,17-HDPAn-6 (∼2%), similar to what is described above for the small scale reaction. In addition, several other dihydroxy compounds (retention time 12–21 min), trihydroxy derivatives (retention time 5–10 min), and oxo and epoxide compounds (retention time 33–36 min) were produced. These were detected, based on characteristic fragmentation patterns as well as UV behavior. For example, the 17-oxo-DPAn-6 showed a UV peak at 280 nm, with a [M-H] of 343 and a diagnostic fragment at 245 along with all other expected fragments. Similarly, the 16,17-epoxide of DPAn-6 exhibited no UV absorbance between 220 and 350 nm and showed the expected [M-H], [M-H]-H₂O, [M-H]-CO₂, and [M-H]-CO₂,-H₂O fragments along with 233, 245, 274, and 261 nm, which are characteristic of an epoxide at the 16,17-position. The total fraction accounted for by all these other compounds was less than 5% of the main products formed. Purification of 17-HDPAn-6 and 10,17-diHDPAn-6 resulted in >98% pure compounds (Fig. 4) that were further characterized and then used for other studies.

FIGURE 3.

LC profile of the gram scale reaction of DPAn-6 with soybean 15-lipoxygenase. DPAn-6 was incubated with soybean lipoxygenase in 0.05 m sodium borate buffer at 4 °C for 30 min. Products were reduced with sodium borohydride, the reaction was acidified, and then products were extracted using DSC-18 material. Products were analyzed by LC/MS as described under “Experimental Procedures.” A, total ion chromatogram; B, UV absorbance at 236 nm; C, UV absorbance at 270 nm.

FIGURE 4.

LC/MS/MS of purified 17-HDPAn-6 and 10,17-HDPAn-6. The reaction mixture represented by Fig. 2 was purified by preparatory HPLC techniques using conditions described under “Experimental Procedures.” Fractions of interest were pooled from several runs and further concentrated and analyzed for purity using LC/MS/MS techniques. A, total ion chromatogram of 17-HDPAn-6 with LC conditions, as described under “Experimental Procedures.” The inset shows the UV spectrum of the main peak; B, mass spectra of the main peak shown in A with main fragments denoted by their origin. The inset shows 17-HDPAn-6 with the expected diagnostic fragments. C, total ion chromatogram of 10,17-HDPAn-6 with an inset showing the UV spectrum of the main peak. D, mass spectra of the main peak shown in C with main fragments denoted by their origin. The inset shows 10,17-HDPAn-6 with the expected diagnostic fragments.

Structural Characterization of 17-HDPAn-6 and 10,17-diHDAPn-6—As described previously, 17-HDPAn-6 showed a UV spectrum with an absorbance maximum at 236 nm. The extinction coefficient in ethanol at room temperature was calculated to be 30,000/m·cm. All MS/MS fragments were consistent with the 17-HDPAn-6 structure, as shown in Fig. 4. Fig. 5 shows the one-dimensional ¹H, one-dimensional ¹³C, and two-dimensional ¹H,¹H COSY spectra of 17-HDPAn-6. The number of peaks, integration, and chemical shifts in both the one-dimensional ¹H and the one-dimensional ¹³C spectra were consistent with the molecular formula C₂₂H₃₄O₃ and were consistent with spectra predicted by Chemdraw and ACDlabs software. The two-dimensional ¹H,¹H COSY unambiguously showed that hydroxylation was at the 17-position. The protons belonging to C₍₁₇₎-H-OH centered at 4.12 ppm produced cross-peaks to C₍₁₆₎-H and C₍₁₈₎-H but not with protons belonging to C₍₁₃₎ or C₍₁₄₎. A two-dimensional ¹H,¹H DQFCOSY, a two-dimensional ¹H,¹³C-HSQC, and a two-dimensional ¹H,¹³C HMBC confirmed all nonoverlapping ¹H and ¹³C assignments and double bond configurations. The double bond configurations were confirmed based on several lines of evidence as outlined below. Butovich et al. (38, 39) have described signature chemical shifts obtained from NMR spectra of several compounds that are indicative of a trans-trans configuration. Specifically, a conjugated (EE)-diene fragment would have produced a quartet with a chemical shift δ of 6.15 to 6.24 ppm, and this signature quartet is clearly absent in all spectra of the 17-HDPAn-6 produced in this study (see Fig. 5) Coupling constant information also suggests that the structure is 17-hydroxydocosa-(4Z,7Z,10Z,13Z,15E)-pentaenoic acid. This is based on the fact that the value of J_13,14 is 11 Hz, indicative of the cis-configuration around this vinyl group, whereas, J_15,16 is ∼15 Hz, again clearly indicating that the configuration around these double bonds is trans. In addition, a comparison of chemical shifts of 17-HDPAn-6 with 17-HDHA produced by Butovich et al. (38) are remarkably similar, indicating similar double bond configurations.

FIGURE 5.

NMR data for 17-HDPAn-6. A, ¹D-¹H spectrum showing ¹H assignments. The C₍₁₇₎-H-OH and the C₍₁₎-COOH protons are not seen due to exchange broadening; thus, the main peaks can be integrated to 32 protons. Any extra peaks are solvent-related. B, a quantitative one-dimensional ¹³C spectrum showing ¹³C assignments. Peaks can be integrated to 22 carbon atoms. Any extra peaks are solvent-related. A tetramethylsilane (TMS) peak was used for referencing. C, two-dimensional ¹H-¹H COSY showing correlations between cross-peaks belonging to adjacent carbon atoms.

10,17-HDPAn-6 produced and purified to 98% homogeneity had a UV spectrum with a λ_max at 270 nm and shoulders at 260 and 280 nm (±1 nm), indicative of a conjugated triene in the molecule (Fig. 4). The extinction coefficient was estimated to be 38,000/m·cm. The parent ion mass with m/z of 361 along with fragments (Fig. 4) confirms hydroxylation at C-10 and C-17 of DPAn-6 and is consistent with the molecular formula C₂₂H₃₄O₄. NMR data for this compound are shown in Fig. 6. The ¹D-¹³C spectrum clearly shows the presence of 22 carbons, with two of the carbons having a chemical shift δ of 71.62 and 71.02 ppm, indicative of hydroxylations at these two carbons. Again, as in the case of 17-HDPAn-6, no peaks with chemical shift δ at 6.12–6.24 ppm were seen that would have indicated the presence of either (11E,13E,15Z)-, (11Z,13E,15E)-, or (11E,13E,15E)-isomers. The ¹H and ¹³C chemical shifts and coupling constants were similar to those of potato lipoxygenase-derived as well soybean lipoxygenase-derived 10,17-dihydroxydocosahexa-(4Z,7Z,11E,13Z,15E, 19Z)-enoic acid for C₍₁₎-C₍₁₈₎ and H₍₁₎-H₍₁₈₎ (38, 39). Coupling constants of 15 Hz for trans-, and 10 Hz for cis-double bond configurations were observed around the triene structure, strongly suggesting that the compound was 10,17-dihydroxydocosa-(4Z,7Z,11E,13Z, 15E)-pentaenoic acid. Although NMR data strongly suggest double bond configurations, as described above, synthetic standards with known double bond configurations will be needed for final confirmation.

FIGURE 6.

NMR data for 10,17-HDPAn-6. A, one-dimensional ¹H spectrum showing ¹H assignments. The C₍₁₀₎-H-OH, C₍₁₇₎-H-OH, and the C₍₁₎-COOH protons are not seen due to exchange broadening; thus, the main peaks can be integrated to 31 protons. Any extra peaks are solvent-related. B, a quantitative ¹D-¹³C spectrum showing ¹³C assignments. Peaks can be integrated to 22 carbon atoms. Any extra peaks are solvent-related. A tetramethylsilane (TMS) peak was used for referencing. C, two-dimensional ¹H-¹H COSY showing correlations between cross-peaks belonging to adjacent carbon atoms.

Chirality—VCD techniques, which are used for ascertaining absolute configurations at chiral centers, were not successful in determination of chirality in the case of 17-HDPAn-6. Therefore, chiral chromatography was used for assigning chirality. Order of elution of R- and S-isomers on chiral columns is the same for similar hydroxylated LC-PUFA compounds when the same chromatographic conditions are used (40, 41). In this case, the 17-HDHA racemate mixture and (17S)-HDHA was used to ascertain the order of elution. The Chirobiotic-T column was not successful in separating isomers, and final separations were successful using the Chiralpak-IA column. As seen in Fig. 7, it is clear that the R-isomer of 17-HDHA elutes before the S-isomer. A similar separation for 17-HDPAn-6 shows that this enzymatically prepared compound comprises 95% S-isomer and 5% R-isomer. Hence, 15-lipoxygenase is primarily pro-S-selective in the case of DHA as well as DPAn-6 at the C-17-position.

FIGURE 7.

Chirality determination of C17 in 17-HDPAn-6. 17-HDHA and 17-HDPAn-6 prepared enzymatically using 15-lipoxygenase were oxidized to the corresponding oxo-compounds and then reduced with sodium borohydride to the corresponding racemic mixtures. Separations were performed on Chiralpak-IA using methanol/water (80:20) as mobile phase at a flow rate of 0.6 ml/min at 25 °C. A, racemic 17-HDHA prepared by oxidation-reduction of enzymatically prepared 17-HDHA; B, (17S)-HDHA from Cayman Chemicals; C, 17-HDHA prepared enzymatically using 15-lipoxygenase; D, racemic 17-HDPAn-6 prepared by oxidation-reduction of enzymatically prepared 17-HDPAn-6; E, enzymatically prepared 17-HDPAn-6 showing the major isomer assigned as S.

Physicochemical Properties and Stability—pK_a values of both compounds were predicted by ACDlabs software to be 4.58. Log D_{pH 7.4} values determined using PBS and octanol partitioning were consistent with those predicted using ACDlabs software (Fig. 8A) and were 3.5 for 17-HDPAn-6 and 2.7 for 10,17-diHDPAn-6. The corresponding calculated log P values were 6.32 and 5.09, respectively. Experimentally, solubility of both compounds in PBS, pH 7.4, was found to be more than 200 μm. Solubilities of reference compounds, as determined in the same assay, were 5.6 μm for diethylstilbestrol, 46.3 μm for haloperidol, 109.2 μm for ketoconazole, 227.3 μm for metoprolol tartrate, 81 μm for phenytoin, 197.8 μm for rifampicin, 19.9 μm for simvastatin, and 0.6 μm for tamoxifen. Predictions using ACDlabs software (Fig. 8B) suggest that 17-HDPAn-6 has a solubility of 1.9 mm and that the more polar 10,17-diHDPAn-6 has a solubility that is about 6-fold higher at 12.2 mm in PBS, pH 7.4. The stability of compounds was tested in PBS and ethanol, as described under “Experimental Procedures.” At 4 °C, there was not more than 4% degradation after 14 days. Also, both compounds were completely stable (0% degradation) for 24 h at room temperature and for 2 h at 37 °C.

FIGURE 8.

Predicted log D and solubility profile of 17-HDPAn-6 and 10,17-diHDPAn-6. Log D and solubility were calculated for 17-HDPAn-6 and 10,17-HDPAn-6 using ACDLabs software. A, a plot of predicted log D at various pH values for 17-HDPAn-6 (○) and 10,17-HDPAn-6 (▴). B, predicted log solubility at various pH values for 17-HDPAn-6 (○) and 10,17-HDPAn-6 (▴).

Anti-inflammatory Activity of DPAn-6 Resolvins in the Air Pouch Model—The murine dorsal air pouch model, mimicking acute inflammation of the synovium, was used to assess anti-inflammatory activity of DPAn-6 resolvins. Resolvins were administered locally by intrapouch injection, and within 5 min, TNFα was injected directly into the pouch to induce the inflammatory response. Exudates were harvested 4 h later, and infiltrating leukocytes were enumerated. Fig. 9 shows the total number of leukocytes in the pouch exudates. As expected, TNFα stimulated an ∼4-fold increase in leukocyte migration into the pouch. Cell migration was reduced to the level of the negative control (no TNFα stimulation) by both 17-HDPAn-6 and 10,17-HDPAn-6, whereas the resolvin control (17S)-HDHA tended to reduce migration, but the reduction was not statistically significant. The major effect of the DPAn-6-derived resolvin compounds was to reduce the proportion of granulocytes and increase the proportion of macrophages (Table 5) in the exudates, suggesting a switch from active immune response (granulocyte-dominated) to resolution (macrophage-dominated) of the inflammatory response. The two DPAn-6 resolvin compounds administered locally at a dose of ∼0.005 mg/kg (on a whole animal basis) were as efficacious as indomethacin administered intraperitoneally at a dose of 2 mg/kg in this study.

FIGURE 9.

Effects of 17-HDPAn-6 and 10,17-HDPAn-6 on leukocyte migration in a murine air pouch model of acute inflammation. The figure shows means (n = 10 animals/group) ± S.D. of the number of leukocytes found in air pouch exudates 4 h after stimulation with TNFα. Test compounds (100 ng) were administered intrapouch immediately before injection of the TNFα into the pouch, and indomethacin was administered intraperitoneally (2 mg/kg) 30 min prior to the TNFα injection. Groups were compared with the TNFα control group by ANOVA with Dunnett's post test. *, p < 0.01 compared with TNFα control.

TABLE 5.

Proportion of granulocytes and macrophages in exudates harvested from murine air pouches

Animals were injected with 100 ng of resolvins followed immediately by 100 ng of TNFα intrapouch. Exudates were harvested 4 h later, and granulocytes and macrophages were quantified after Diff-Quik staining on slides. Means ± S.D.

Group	Granulocytes	Macrophages
	%	%
TNFα control	81.8 ± 5.4	18.2 ± 5.4
No TNFα control	61.2 ± 19.4a	38.8 ± 22.3a
17-HDHA	76.3 ± 19.5	23.7 ± 19.5
17-HDPAn-6	67.4 ± 4.1a	32.6 ± 4.1a
10,17-HDPAn-6	66.8 ± 15b	33.2 ± 15.0b
Indomethacin	70.9 ± 16.1	29.1 ± 16.1

^ap < 0.01, significantly different than TNFα control by Student's t test.

^bp < 0.05, significantly different than TNFα control by Student's t test.

Rat Hind Paw Edema Assay—The two DPAn-6 resolvins were tested in a rat carrageenan-induced hind paw edema model to confirm their anti-inflammatory activity in a second animal species and in a second in vivo model of acute inflammation. This also served to explore efficacy following systemic rather than local administration. The intravenous doses in this study were 50-fold higher than in the murine air pouch study to account for the larger species size and to account for systemic rather than local administration. Oral doses were an additional 5-fold higher than intravenous doses to help compensate for possibly lower bioavailability. Results, shown in Fig. 10, indicate that both 17-HDPAn-6 and 10,17-H DPAn-6 reduced edema volume following intravenous administration and resulted in ∼20–25% reduction in paw edema at the 2, 4, and 6 h time points, demonstrating activity of these compounds for at least 6 h after intravenous administration. Potency was similar to the known resolvin 17-HDHA at 4 and 6 h, although the 17-HDHA appeared more potent at the 2 h time point. There was no effect of any compound, including indomethacin, at the 24 h time point. The 10,17-HDPAn-6 was active orally, resulting in similar reductions in edema as with the intravenous dosing of this compound. The 17-HDPAn-6 did not show activity at this dose following oral administration in this study.

FIGURE 10.

Kinetics of effect of intravenously and orally administered DPAn-6 resolvins on carrageenan-induced hind paw edema in rats. Animals were treated with 5 μg of resolvins intravenously 2 min prior to carrageenan challenge (A) and 25 μg of resolvins orally 30 min prior to carrageenan challenge (B). Vehicle (○) was administered intravenously 2 min prior to carrageenan challenge, and indomethacin (×) was dosed at 5 mg/kg intraperitoneally 30 min prior to carrageenan challenge in both A and B. Resolvins shown are 17-HDPAn-6 (▪), 10,17-HDPAn-6 (•), and 17-HDHA (▴) in both A and B. Means (n = 8 animals/group) ± S.D. of edema volume measured 2, 4, 6, or 24 h post-carrageenan challenge are shown. **, significantly different from vehicle control by ANOVA with Dunnett's post test, p < 0.01.

A second hind paw edema study was conducted to explore the dose-response relationship of orally administered DPAn-6 resolvins. Data from this study are summarized in Fig. 11.

FIGURE 11.

Dose effects of 17-HDPAn-6 and 10,17-HDPAn-6 on carrageenan-induced rat paw edema. All compounds were administered by oral gavage 30 min before carrageenan challenge, and paw edema was measured 2, 4, or 6 h after carrageenan challenge. The figure shows mean paw edema volume (n = 6 animals/group) ± S.D. after administration of various doses of 17-HDPAn-6 (A) or 10,17-HDPAn-6 (B). Groups represented are vehicle (○), 2.5 μg (•), 25 μg (▪), and 250 μg (▴). Dose effects of compounds at the 4 h time point are shown in C. Resolvin treatment groups were compared with vehicle control. *, p < 0.05; **, p < 0.01 by ANOVA with Dunnett's post test.

The responses in this dose-response assay reproduced the results from the first hind paw assay in that the 25-μg dose of 10,17-HDPAn-6 produced a significant ∼20% reduction in paw edema, whereas the same dose of the 17-HDPAn-6 had no effect. The response to the resolvin 17-HDHA at 25 μg resembled that of 17-HDPAn-6 with no reduction in edema volume, whereas the indomethacin reduced swelling by about 30%. Although there was no effect of 17-HDPAn-6 at the middle (25-μg) dose, the low (2.5-μg) and high (250-μg) doses reduced edema at the 4 h time point (Fig. 11C). The high dose of 17-HDPAn-6 was most effective and consistently reduced edema at all time points (Fig 11A). On the other hand, the middle dose of 10,17-HDPAn-6 was most effective at reducing edema at all time points (Fig. 11B). The low and high doses of this compound reduced edema at the 2 h but not later time points. Both DPAn-6 resolvin compounds resulted in nonmonotonic dose responses (Fig. 11C).

DISCUSSION

Lipoxygenases catalyze the dioxygenation of LC-PUFAs containing cis,cis-1,4-pentadiene moieties (26). The primary site of oxidation in arachidonic acid is commonly used in naming these enzymes. For example, 15-lipoxygenase catalyzes the production of 15-hydroperoxyarachidonic acid from ARA (42). In the current study, 15-lipoxygenase, 12-lipoxygenase, and 5-lipoxygenase acted on DHA, DPAn-6, and DPAn-3 with positional preferences for dioxygenation. 15-Lipoxygenase, primarily oxygenated the C22 substrates at the C-17 or ω-6 position, as would be expected from previous studies with these enzymes (3, 4, 43). In addition, 10,17- and 7,17-dihydroxyoxylipins were also obtained from DPAn-6 and DPAn-3. These compounds are analogous in structure to the potent anti-inflammatory 17-hydroxy series docosanoids described by Hong et al. (3). Other lipoxygenases like 12-lipoxygenase primarily oxygenated DHA, DPAn-3, and DPAn-6 at the C-14/ω-9-position and 5-lipoxygenase at the C-10-position, consistent with what was expected (4, 44–46). Like DHA, other LC-PUFAs, such as DPAn-6 and DPAn-3, and lipoxygenases are found in varying amounts in blood and several tissues, suggesting that these or other DPA-derived resolvins could be produced physiologically and act as bioactive mediators (47–52). However, actual biological production of these DPAn-6-derived resolvins remains to be investigated. Comparison of substrate reactions with 15-lipoxygenase at 0 °C showed that DPAn-6 was completely converted to products, whereas DHA and DPAn-3 were not, thus indicating that at such low temperatures, DPAn-6 reacts with 15-lipoxygenase in a facile manner. When enzyme kinetics behavior was probed at higher temperatures (22 °C) in the presence of a surfactant, DPAn-3 exhibited the highest V_max. However, DPAn-6 proved to be the substrate that was converted most efficiently to products. The 15-lipoxygenase-PUFA reaction is a complex one, involving several phenomena, such as substrate inhibition, product inhibition, product activation, substrate deactivation from decoupling of the radical substrate prior to oxygen insertion, low oxygen concentration, and enzyme thermodynamic inactivation. Additionally, the substrates exhibit tendencies toward micelle formation on which 15-lipoxygenase does not act effectively (33). The critical micelle concentration for linoleic acid at pH 9.0 and 20 °C is 150 μm (53), and the critical micelle concentration for DHA at pH 8.8 and 25 °C is 60–90 μm (54). Although some of the factors mentioned above were taken into consideration while determining kinetic constants, it was not feasible to evaluate all of them as is true for almost all studies evaluating the 15-lipoxygenase-PUFA reaction. Therefore, the experiments reported here are useful for comparing the efficiency with which soybean 15-lipoxygenase converts DHA, DPAn-3, and DPAn-6 to products, but comparison with other studies remains difficult, due to the large variability of conditions reported in the literature. Both DHA and DPAn-3 are ω-3 fatty acids, and the different kinetic behavior observed for DPAn-6 could be due to the fact that it is an ω-6 fatty acid. This could affect the affinity of the substrate to the enzyme at the reactive site as well as at the site responsible for substrate inhibition. In the absence of precise structural information regarding details of binding interactions between the substrate and the enzyme, it is difficult to pinpoint the precise origin of the differences observed in the kinetic activities.

One of the main reasons for focusing our attention on DPAn-6-derived oxylipins stems from the fact that it is a constituent of our Martek-DHA-S™ oil (27). Studies have shown that in a carrageenan-induced rat hind paw edema model, this Martek-DHA-S™ oil performed better at reducing paw volume than the Martek-DHA-T™ oil.⁵ The major difference between the two oils is that in addition to DHA, the DHA-S™ oil contains DPAn-6. We thus reasoned that DPAn-6 contributes to the anti-inflammatory activity of this oil and possibly does this via resolvin-like mediators. Since most resolvins are oxylipins produced by action of lipoxygenases on LC-PUFAs, we examined if DPAn-6 was a substrate for lipoxygenases and characterized the products of these reactions. The major products of the most efficient reaction, 17-HDPAn-6 and 10,17-HDPAn-6, were extensively characterized and assessed for their “oral drugability” characteristics and further examined for their anti-inflammatory potential. Chemical characterization of these resolvins involved determination of double bond configurations and chirality at asymmetric carbon atoms, since these are important chemical features that affect bioactivity. For example, in the neuroprotectin D1 family, the major isomer produced by human leukocytes, characterized as (10R,17S)-dihydroxy-(4Z,7Z, 11E,13E,15Z,19Z)-hexaenoic acid, was found to be more potent than (10S,17S)-dihydroxy-(4Z,7Z,11E,13Z,15E,19Z)-hexaenoic acid, in reducing neutrophil migration in a murine zymosan-induced peritonitis model (1). It should be pointed out that ¹⁸O₂ incorporation experiments have shown that the former is generated via an epoxide intermediate, whereas the latter is produced primarily via a double lipoxygenation reaction. Although another isomer, (10S,17S)-dihydroxy-(4Z,7Z, 11E,13E,15Z,19Z)-hexaenoic acid, showed substantial activity in the same study, it was not found to be biologically relevant. In our study, configurations of double bonds in 17-HDPAn-6 and 10,17-HDPAn-6 as elucidated by high resolution NMR were the same as analogous DHA products described by Butovich et al. (38, 39). In particular, 10,17-HDPAn-6 has double bond geometry that is different from the main neuroprotectin D1 isomer found in human neutrophils and is in effect consistent with a double lipoxygenation product, the major isomer in murine exudates and a minor isomer in human leukocytes, as described by Serhan et al. (1, 39).

Soybean lipoxygenase is known to primarily produce LC-PUFA hydroperoxides with an S-configuration at the chiral ω-6 carbon. This has been seen with several substrates, such as docosahexaenoic acid, arachidonic acid, and linoleic acid, when intact soybean lipoxygenase (Type 1-B) was used (38, 55). Serhan et al. have demonstrated that (17S)-HDHA is the main product of the reaction of DHA with soybean lipoxygenase and that in order to obtain (17R)-hydroxylated compounds of DHA, one needs to use acetylated COX-2 (3, 4). Our chiral chromatography data clearly show that DPAn-6 is converted by soybean lipoxygenase to (17S)-HDPAn-6. Commonly used techniques, such as VCD, that are used to determine configurations at chiral centers did not work for 17-HDPAn-6. This is primarily because this class of molecules is highly flexible, resulting in multiple conformations, thus complicating the assignment of VCD/infrared spectra to a single conformer. Formation of 10,17-HDPAn-6 as described before seems to occur through a double lipoxygenation mechanism, and we thus believe that this dihydroxy compound possesses the S-configuration at C-17 and C-10, a situation analogous to what is seen for 10,17-HDHA described by Butovich et al. (38, 39).

Oral drugability of these two DPAn-6 derived resolvins can be evaluated using the commonly used Lipinski's rule of five (56–58). Based on this rule, a good orally administered drug candidate should have the following features: molecular weight less than 500, fewer than five hydrogen bond donors, fewer than 10 hydrogen bond acceptors, and a partition coefficient, log P less than 5 (58). The two compounds, 17-HDPAn-6 and 10,17-HDPAn-6, satisfy all of the criteria except that their log P values are 6.32 and 5.09, respectively. Despite this apparently high hydrophobicity, the solubility of these compounds was better than that of a large number of other commonly used drugs, such as simvastatin, tamoxifen, haloperidol, and phenytoin. Nevertheless, the log P values of both DPAn-6-derived resolvins could be easily changed with suitable derivitization.

17-HDPAn-6 and 10,17-HDPAn-6 used in efficacy testing were produced biogenically using soybean 15-lipoxygenase, and each was highly purified (>98% purity), ensuring that the responses seen in the animal studies were indeed due to these specific DPAn-6 resolvins. In the air pouch model of acute inflammation, both 17-HDPAn-6 and 10,17-HDPAn-6 were highly efficacious at reducing leukocyte migration into inflamed pouches. The 100-ng DPAn-6 resolvin dose, chosen to match doses used by Serhan and colleagues (3, 4) with DHA-resolvins in the same model, reduced leukocyte migration by 70–80% and had effects similar to that reported for (17S)-series DHA-derived resolvins, which reduced neutrophil migration into air pouches. In our experiment, the 17-HDPAn-6 resolvins were more efficacious than the same dose of 17S-HDHA, although different stereoisomers of the resolvins may have different potencies (1). Importantly, both DPAn-6 resolvins reduced leukocyte migration to the base-line levels seen in animals not receiving TNFα, and both compounds were as efficacious as the indomethacin administered intraperitoneally but at a 400 times higher dose (on a whole animal basis). Interestingly, both compounds proportionately reduced granulocyte and increased macrophage numbers. This is consistent with a proresolution mechanism described by Gilroy et al. (59), wherein neutrophil migration is reduced and macrophages move in to phagocytose the apoptotic neutrophils. Further research is required to understand the specific mechanism by which the DPAn-6 resolvins mediate their response, but in the air pouch model of acute inflammation, they behave similarly to the structurally analogous DHA-derived resolvins.

The hind paw edema model is a widely used model of acute inflammation and has been recently used to test the analgesic effects of lipoxins (60). In our studies, paw inflammation was induced by injecting cross-linked carrageenan. Despite this strong stimulus, both 17-HDPAn-6 and 10,17-HDPAn-6 significantly reduced paw edema by 20–25% when administered systemically using a 5-μg intravenous dose (∼0.025 mg/kg). This response was similar to the response to biogenically synthesized (17S)-HDHA, also administered intravenously. Both compounds also showed efficacy at early time points (2 h post-carrageenan) when administered orally at the lowest dose (2.5 μg) tested. However, both compounds exhibited nonmonotonic dose-response behavior with the oral administration route. The 17-HDPAn-6, for example, was not active at the middle 25-μg dose (similar to its DHA analog 17-HDHA) but had activity at the low and high doses. The most effective doses across all time points, resulting in about 25% reduction in edema, were 250 μg for 17-HDPAn-6 and 25 μg for 10,17-HDPAn-6. These observations suggest that either 10,17-HDPAn-6 is more potent than 17-HDPAn-6, that bioactive metabolites of 10,17-HDPAn-6 contribute to its response, or that early exposure to 10,17-HDPAn-6 may result in extended effects by, for example, altering gene expression patterns. Although we do not know the reason for the nonmonotonic dose-response behavior observed here, possible explanations include dose-dependent metabolism in vivo, interaction with multiple receptor systems, or mixed agonist/antagonist activity profiles (61, 62). Based on strong similarities to the DHA-derived resolvins, in both their synthesis by lipoxygenases to structurally analogous compounds and their comparable activity at similar doses in acute inflammation models, these DPAn-6-derived oxylipins appear to function like resolvins, which act in a proresolution fashion to down-regulate inflammation and can thus be categorized as DPAn-6-derived resolvins. Further studies will be required to determine whether persistent, inappropriate disease-associated inflammation, such as that found in cardiovascular disease or autoimmune diseases, can be modified by providing either DPAn-6 fatty acid precursors or by direct administration of these DPAn-6-derived resolvins or analogs thereof. Although a large number of studies need to be done to assess the bioefficacy and therapeutic profile of both 17-HDPAn-6 and 10,17-HDPAn-6, these could potentially represent a new class of therapeutic resolvins.