| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Center for Experimental Therapeutics and Reperfusion Injury, Department of Anesthesiology, Perioperative, and Pain Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA 02115; and
Department of Chemistry and the Loker Hydrocarbon Research Institute, University of Southern California, Los Angeles, CA 90089
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
15-trans-PD1 (isomer III), 10S,17S-dihydroxy-docosa-4Z,7Z,11E,13Z,15E,19Z-hexaenoic acid (isomer IV), and a double dioxygenation product 10S,17S-dihydroxy-docosa-4Z,7Z,11E,13Z,15E,19Z-hexaenoic acid (isomer I), present in exudates. 18O2 labeling showed that 10S,17S-diHDHA (isomer I) carried 18O in the carbon-10 position alcohol, indicating sequential lipoxygenation, whereas PD1 formation proceeded via an epoxide. PD1 at 10 nM attenuated (
50%) human neutrophil transmigration, whereas 15-trans-PD1 was essentially inactive. PD1 was a potent regulator of polymorphonuclear leukocyte (PMN) infiltration (40% at 1 ng/mouse) in peritonitis. The rank order at 1- to 10-ng dose was PD1 
PD1 methyl ester >> 15-trans-PD1 > 10S,17S-diHDHA (isomer I). 10S,17S-dihydroxy-docosa-4Z,7Z,11E,13E,15Z,19Z-hexaenoic acid (isomer VI) proved 
PD1 in blocking PMN infiltration, but was not a major product of leukocytes. PD1 also reduced PMN infiltration after initiation (2 h) of inflammation and was additive with resolvin E1. These results indicate that PD1 is a potent stereoselective anti-inflammatory molecule. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Further studies on the endogenous mechanisms of anti-inflammation using a murine model of spontaneous resolution demonstrated, for the first time, that resolution is an active biochemical process that involves the generation of specific new families of lipid mediators (for recent reviews, see Refs.16 and 17). During spontaneous resolution, cell-cell interactions and transcellular biosynthesis lead to the production of these new families of potent bioactive lipid mediators from omega-3 essential fatty acid precursors, and were termed resolvins (resolution phase interaction products derived from docosahexaenoic acid (DHA, C22:6) and eicosapentaenoic acid (EPA, C20:5)) and protectins (docosatrienes derived from DHA; Refs.1 , 3 , and 18 and recently reviewed in Ref.19). These novel di- and tri-hydroxy-containing products from EPA and DHA that are generated by previously unrecognized enzymatic pathways display potent anti-inflammatory andimmunoregulatory actions in vitro and in vivo in murine models of acute inflammatory actions (1, 3, 18).
In 1929, the omega-3 polyunsaturated fatty acids were assigned essential roles because their exclusion from the diet gave rise to a new form of deficiency disease (20). Many recent reports document the importance of fish oil (omega-3) fatty acids EPA and DHA in human diseases associated with inflammation. In particular, omega-3 DHA and EPA are protective in inflammatory bowel disease and colitis (21), cardiovascular disease (22, 23, 24, 25), and Alzheimer’s disease (26). However, the molecular mechanisms responsible for these well-documented beneficial actions of omega-3 fatty acids remain an important challenge. DHA is enriched in neural tissues, where it appears to play functional as well as structural roles (27, 28). Along these lines, results from earlier studies indicated that DHA was enzymatically converted to products coined docosanoids, which might be linked to retinal protection (29) and neuronal function (30). The structures of the molecules involved, however, were not established.
Human whole blood isolated leukocytes, and glial cells enzymatically convert DHA to 17S-hydroxy-containing docosatrienes and 17S-series resolvins (1, 3). The novel 10,17S-docosatriene, first identified in Ref.3 and its basic structure established, displayed potent anti-inflammatory actions, i.e., reducing PMN numbers in exudates in vivo and down-regulating production of proinflammatory cytokines by glial cells in vitro (1). During the resolution phase of peritonitis, unesterified DHA levels increase within exudates, and 10,17S-docosatriene is generated within the resolving exudates, where it appears to promote catabasis, or the return to homeostasis, by shortening the resolution interval (31). Of special interest, this DHA-derived 10,17S-docosatriene is generated in vivo during strokes in murine tissues and limits the entry of leukocytes into the area of neural damage, reducing the magnitude of tissue injury (32). In collaboration with Bazan and colleagues (2), we found that 10,17S-docosatriene is neuroprotective in retinal pigmented cells and introduced the term neuroprotectin D1 (NPD1) for this potent compound, which accumulates in the ipsilateral hemisphere of the brain following focal ischemia (33).
Recent results indicate that NPD1 is formed from DHA in cornea in a lipoxygenase (LO)-dependent fashion to protect from thermal injury as well as promote wound healing (34). It is noteworthy that NPD1, resolvin D1, and resolvin D5 are all produced by trout brain cells from endogenous DHA, suggesting that the structures of these DHA-derived mediators are conserved from fish to humans (35). Together, these recent findings underscore the need to establish the complete stereochemistry of endogenous biologically active 10,17S-docosatriene, namely its carbon-10 position alcohol chirality and double bond geometry of its conjugated triene system. In recognition of its wide scope of formation and actions, protectin D1/neuroprotectin D1 (PD1) is used to denote the structure of this chemical mediator, and the prefix neuro before protectin D1 is used to note its tissue origin and address. In this study, we report the complete stereochemistry of protectin D1 and its related natural isomers (i.e., 15-trans-PD1) as well as their anti-inflammatory properties.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Zymosan A, soybean LO (fraction V), and calcium ionophore, A-23187, were purchased from Sigma-Aldrich. DHA and 5-LO from potato (pt5LO) were obtained from Cayman Chemical. Additional materials used in liquid chromatography (LC)-UV-tandem mass spectrometry (MS-MS) analyses were obtained from vendors reported in Refs.1 and 3 . 18O2 isotope was purchased from Cambridge Isotopes.
Isolation, LC-MS-MS, and gas chromatography (GC)-MS analyses
Incubations were extracted with deuterium-labeled internal standard (prostaglandin E2) (Cayman Chemicals) for LC-MS-MS analysis using a Finnigan LCQ LC ion trap tandem mass spectrometer equipped with a LUNA C18-2 (150 x 2 mm 5 µm) column and a rapid spectra scanning UV diode array detector using mobile phase (methanol:water:acetate at 65:35:0.01) with a 0.2 ml/min flow rate that monitored UV absorbance 0.1 min before samples entered the MS-MS. The scan acquisition rates were 11/min for MS-MS and 60/min for UV, which give rise to a lag interval in retention times that was corrected in the results presented for each molecule. All intact cell incubations and in vivo exudates were stopped with 2 vol cold methanol and kept at 20°C for >30 min. Samples were extracted using C18 solid phase extraction and further analyzed using GC-MS, using a Hewlett-Packard 6890 with a HP 5973 mass detector (Table I), and LC-MS-MS. Detailed procedures for isolation, quantitation, and structural determination of these DHA and related lipid-derived mediators, as reported recently in Refs.1 and 36 , were used in this study for elucidation of new products. Biogenic synthesis of some of the DHA-derived products were performed using isolated enzymes, i.e., 5-LOX from potato and 15-LO were each incubated in tandem sequential reactions (see Refs.1, 3, 37 and 38) with either DHA, 17S-hydroxy-DHA, or 17S-hydro(peroxy)-DHA (17S-H(p)DHA) to produce the compounds in quantities suitable for isolation and incubation with cells and tissues as well as confirmation of physical properties and assignment of biological actions. Incubations in an 18O2-enriched atmosphere were performed and analyzed as in Ref.39 .
|
1H NMR (400 MHz, MeOH-d4) were as follows: 
6.52 (dd, J = 14.1 Hz and 11.8 Hz, 1H), 6.26 (m, 2H), 6.07 (dd, J = 11.1 Hz and 11.1 Hz, 1H), 5.50–5.28 (m, 7H), 4.90 (s, 2H), 5.50–5.60 (m, 2H), 4.55 (m, 1H), 4.14 (m, 1H), 3.65 (s, 3H), 2.82 (m, 2H), 2.40–2.13 (m, 8H), 2.06 (m, 2H), 2.07 (m, 2H), 0.96 (t, J = 7.5 Hz, 3H).
13C NMR (125 Hz, MeOH-d4) were as follows: 174.93, 137.59, 134.56, 134.47, 134.35, 131.01, 130.52, 130.17, 129.92, 128.57, 128.52, 126.14, 124.89, 72.60, 68.18, 36.00, 35.97, 34.45, 26.30, 23.43, 21.312, 14.20. The total organic synthesis of PD1 and related isomers will be reported separately (3).
Incubations with human PMN and whole murine brain
Human venous whole blood (10 ml) was collected into heparin (0.01 U/ml) via venipuncture from healthy volunteers (who declined taking medication for 2 wk before donation; Brigham and Women’s Hospital protocol 88-02642). Human PMN were freshly isolated from whole blood using Ficoll gradient and enumerated as in described in Ref.40 . PMN (30 x 106 cells/ml) were exposed to ionophore A23187 (5 µM) with either DHA or 17S-H(p)DHA (15 µg/ml). Cell suspensions were incubated for 30 min at 37°C in a covered water bath. Lipidomics and lipid mediator profile analyses were conducted for DHA-derived products, resolvins, and docosatrienes, as described in Refs.1 , 3 , 18 , and 36 . Whole murine brains (Charles River Laboratories) were excised, and then homogenized in PBS (minus Mg2+ and Ca2+) (Cambrex Bioscience). Brain homogenates were washed once (800 rpm, 4°C, 5 min) in PBS, and then suspended in 1 ml of PBS minus divalent cations. Homogenates were placed in the incubator (5 min) with an atmosphere of 5% CO2 at 37°C, and calcium ionophore A23187 (5 µM) or DHA (30 µM) was added. Brain homogenates were then placed in the incubator for 30 min, with an atmosphere of 5% CO2 (37°C). Incubations were stopped with 2 vol of ice-cold MeOH, and, after 30 min at 4°C, the suspensions were pelleted, and the supernatants were extracted using solid phase C18 cartridges (Alltech Associates). Material eluted with methyl formate was taken to dryness using a stream of nitrogen gas, and resulting material was further analyzed.
Human neutrophil transmigration
PMN were freshly isolated from whole blood obtained by venipuncture from healthy human donors (who denied taking medication for 2 wk before donation; Brigham and Women’s Hospital (Boston, MA), protocol no. 88-02642) and anticoagulated with acid citrate dextrose as described in Refs.40 and 41 . Briefly, plasma and mononuclear cells were removed by aspiration from the buffy coat after centrifugation (400 x g; 20 min) at room temperature. Histopaque (density 1.077) was obtained from Sigma-Aldrich. RBCs were sedimented using 2% gelatin, and residual RBCs were removed by lysis in ice-cold NH4Cl buffer. The cell suspensions were >90% PMN, as determined by light microscopic evaluation. PMN were suspended at 5 x 107 cells/ml in HBSS with 10 mM HEPES (pH 7.4) and without Ca2+ or Mg2+ (Sigma-Aldrich). PMN were used within 2 h of their isolation.
Human microvascular endothelial cells (HMEC; a gift from Francisco Candal, Centers for Disease Control, Atlanta, GA) were obtained as primary cultures. For preparation of experimental HMEC monolayers, confluent endothelial cells were grown on 0.33-cm2 ring-supported polycarbonate filters (5-µm pore size; Costar) in the apical-to-basolateral direction. Cells were grown for 1 wk before transmigration experiments.
Transmigration was conducted essentially as described in Ref.41 . PMN were incubated with either vehicle-containing buffer or compound for 15 min at 37°C. A chemotactic gradient was established by placing HMEC inserts that had been washed in HBSS with Ca2+ and Mg2+ (denoted+/+) in 10–8 M LTB4 in the lower chambers. Neutrophils (106 cells) were added to 50 µl of HBSS+/+ in the upper chambers, and cells were incubated at 37°C for 90 min. Transmigrated PMN were quantified by assessing the PMN azurophilic marker myeloperoxidase and a calibration curve. PMN were lysed by the addition of Triton X-100 to a final concentration of 0.5%. The samples were acidified with citrate buffer (final concentration 100 mM; pH 4.2). An aliquot of each sample was added to an equal volume of ABTS solution (1 mM ABTS, 0.03% H2O2, 100 mM sodium citrate buffer (pH 4.2)) in a 96-well plate. The resulting color was monitored using a plate reader at 405 nm, and a calibration curve was used to determine cell number.
Acute inflammatory exudates: murine peritonitis
Peritonitis was conducted using 6- to 8-wk-old FVB male mice (Charles River Laboratories) that were fed laboratory Rodent Diet 5001 (Purina Mills) that were anesthetized with isoflurane, and compounds to be tested were administered i.p. Zymosan A in 1 ml of saline (1 mg/ml) was injected 1–1.5 min later in the peritoneum. Each compound tested or vehicle alone was suspended in 1 µl of ethanol and mixed in sterile saline (120 µl). After i.p. injections (either 2 or 4 h of acute inflammation), the mice were sacrificed in accordance with the Harvard Medical Area Standing Committee on Animals (protocol no. 02570), and peritoneal lavages were collected rapidly and placed in an ice bath (4°C) for enumeration, differential counts, and further analysis.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
The DHA-derived 10,17-dihydroxy conjugated triene-containing product PD1 is generated by several human cell types, murine exudates, skin, and brain tissues (1, 2, 3, 32), as well as isolated fish tissues, indicating that it is a conserved structure in evolution (35). PD1 displays potent protective and anti-inflammatory actions (1, 2, 18, 32). To determine the complete stereochemical assignment of PD1 (i.e., the chirality at C10 and the geometry of the triene), we directly compared the physical and biological properties of DHA-derived PD1 and related 10,17 dihydroxy-docosatriene stereoisomers to those prepared by total organic and biogenic synthesis. These included the following, as illustrated in Figs. 1 and 2: compound I, 10S,17S-dihydroxy-docosa-4Z,7Z,11E,13Z,15E,19Z-hexaenoic acid; compound II, 10R,17S-dihydroxy-docosa-4Z,7Z,11E,13E,15Z,19Z-hexaenoic acid, and compound V, 10S,17R-dihydroxy-docosa-4Z.7Z,11E,13E,15E,19Z-hexaenoic acid, coelute in this HPLC system; compound III, 10R,17S-dihydroxy-docosa-4Z,7Z,11E,13Z,15E,19Z-hexaenoic acid, compound IV, 10R,17S-dihydroxy-docosa-4Z,7Z,11E,13Z,15Z,19Z-hexaenoic acid; and compound VI, 10S,17S-dihydroxy-docosa-4Z,7Z,11E,13E,15Z,19Z-hexaenoic acid. The total synthesis of these will be reported elsewhere (N. A. Petasis, R. Yang, and C. N. Serhan, manuscript in preparation). Of interest, compounds I and VI coeluted in this system, as did both compounds II and V (Figs. 1 and 2). Bioactive PD1 was generated by isolated human cells and murine brain tissue and during inflammation in vivo. Fig. 1 shows representative chromatographic profiles for PD1 generated by isolated human neutrophils incubated with DHA that was separated into several positional and geometric isomers (Fig. 1A, top panel). The main positional isomer of 10,17S-docosatriene was 7,17-diHDHA (denoted as resolvin D5), as documented earlier (1, 3), and a double dioxygenation product (37). Also, a representative profile of products is given for those obtained from murine inflammatory exudates (Fig. 1A, middle panel) and neural tissues (data not shown). A direct comparison of these materials is shown in Fig. 1A, along with the profile obtained for synthetic materials (Fig. 1A, bottom panel) and chromatograms recorded by MS-MS (Fig. 1, right side) and UV at 270 nm (Fig. 1A, left side).
|
|
The chirality of the alcohols and double bond geometry of the triene were systematically addressed. Fig. 1B shows the MS-MS spectrum of PD1 obtained from murine peritonitis (4 h) generated in vivo upon challenge with zymosan A. Fig. 1C shows the mass spectrum recorded using the same instrument settings and conditions with synthetic 10R,17S-diHDHA (compound II; Figs. 1 and 2). To obtain additional evidence for matching, GC-MS analyses were performed. Fig. 1D illustrates a representative mass spectrum and prominent ions obtained with GC-MS for PD1 treated with diazomethane and subsequently converted to its corresponding trimethylsilyl derivative. Hence, chromatographic behavior and prominent ions in two mass spectrometry systems (LC-MS-MS and GC-MS), together with biological activity (Figs. 5 and 6), permitted criteria for assignment of the physical properties of PD1 and related isomers. Because the parent and daughter ions were the same for each isomer, retention time in two chromatographic systems and bioactivity were key requirements for assigning the stereochemistry of the endogenous PD1 (vide infra).
|
|
|
maxMeOH for a conjugated triene chromophore with a maxMeOH at 270 nm, with shoulders at 260 and 282 nm (±2 nm). Each isomer gave a specific maxMeOH, which appeared to reflect the geometry of the double bond system. For example, the 15-trans isomer in the conjugated triene portion of 10,17-diHDHA gave an UV maxMeOH of 269 nm (Table I). Only one of these products (compound II) matched the chromatographic behavior using both LC-MS and GC-MS as well as biological activity. As expected, each of the major prominent ions for these isomers in both LC-MS-MS and GC-MS were essentially identical (i.e., daughter and parent ions were essentially the same for each).
The main materials isolated from human PMN coeluted with compound II and, in some preparations, a trace amount of compound IV. Compound IV differs from compound II in its triene configuration, which is 11E,13Z,15E. This change in two double bonds was unlikely in view of the earlier identification of alcohol trapping products (1). Also, compound IV was not observed in all PMN incubations, which might reflect some degree of donor variation. A second representative profile from another human donor is illustrated in Fig. 1A, top right panel, at m/z 260.8 to 261.8 of the MS-MS for M-1 at m/z = 359. As noted for this ion plot and donor, compound II is the most abundant, and compound IV is not present. Also, compound I is clearly present along with an unknown material denoted with an asterisk. In Fig. 1A, top right, the retention time of isomer III is plotted in gray for direct comparison. Only trace amounts of compound III, the 15-trans isomer of PD1, were routinely identified. The appearance of this 15-trans isomer with its triene in the all-trans configuration likely reflects workup-induced isomerization at the 15 position, which may account for its varied presence in LC-MS-MS-based analysis. This is in addition to 15-trans-PD1 formation via nonenzymatic hydrolysis of the proposed epoxide intermediate (Fig. 4B).
|
|
To test the role of sequential LO actions in the proposed mechanism of PD1 formation (compound II; Fig. 2) and its isomer 10S,17S-diHDHA (compound I), incubations were conducted in an atmosphere enriched in isotope 18O2 with 17S-H(p)DHA as substrate and isolated pt5LO (see Materials and Methods). Note that compounds I and II differ in both chirality at carbon-10 as well as geometry of their respective triene configurations (Fig. 2). After extraction and isolation, the product profiles, GC-MS and LC-MS-MS results indicated that 18O was incorporated in the carbon-10 position in 10S,17S-diHDHA (Fig. 4). Chromatographic separation of 10S,17S-diHDHA (Fig. 4) gave prominent ions with MS-MS (Fig. 4A), indicating on average >75% incorporation of 18O originating from molecular oxygen in the carbon-10 position with a range of 51.4 to 91.8% increase in diagnostic ions. Because these enzymes use molecular oxygen as a substrate, it is not possible, under these conditions, to completely replace enzyme-associated 16O for the 18O isotope as calculated earlier for LXA4 in Refs.39 and 44 . The extent of 18O present in diagnostic ions was determined for m/z 181/183, 261/263, 289/291, 297/299, 315/317, 323/325, 341/343, and 359/361, and the ratio of 16O:18O was calculated from ion intensities and averaged. These results indicate that 10S,17S-diHDHA can be produced via double lipoxygenation.
Results from matching studies indicated that the double bond geometry for the conjugated triene portion of this molecule was in the trans,cis,trans configuration (matching compound I; Fig. 2). Hence, double dioxygenation to form 10S,17S-diHDHA was also a mechanism to generate this compound in vivo, because it is a prominent product in murine exudates from peritonitis, and, to some extent, present in suspensions of human leukocytes incubated with DHA (Fig. 1 and Ref.1) and murine brain (3, 18), as well as trout leukocytes and brain (35). Fig. 4B outlines the proposed scheme and proposed role for double dioxygenation and its products 10S,17S-diHDHA and 7S,17S-diHDHA. The double bond geometry in the conjugated triene portion of the molecule (trans,cis,trans) is consistent with oxygenation, using molecular oxygen with two sequential lipoxygenation steps. Given the biological actions, chromatographic and physical properties of PD1 as well as the results from epoxide trapping experiments with human PMN, and the isolation of two vicinal diol 16,17S-dihydroxy-docosatrienes as minor products (1), it is likely that, once a 16,17-epoxide-containing intermediate is generated in situ (as illustrated in Fig. 4), an enzymatic reaction is needed to efficiently produce PD1 carrying the 10R,17S-dihydroxy-trans,trans,cis configuration arising from an epoxide intermediate as depicted in Fig. 4B.
Anti-inflammatory actions of PD1
As indicated above, the complete stereochemical assignment for synthetic PD1 also relied on determining biological action of the related isomers. Earlier results indicated that PD1’s anti-inflammatory properties were comprised of blocking leukocyte infiltration in murine systems (1, 3, 32, 37). Results in Fig. 5 show that synthetic PD1 reduced PMN transmigration in response to LTB4. Amounts as small as 1.0 nM gave 30% inhibition. The 15-trans isomer of PD1, where the conjugated triene portion of the molecule was in the trans configuration, did not block PMN transmigration in vitro. Although PD1 is a potent inhibitor in neutrophil transmigration, the degree of inhibition observed with monolayers of HMECs and human neutrophils from >5 separate donors did not achieve values greater than IC50 in each experiment.
These experiments with transmigration were conducted in parallel with murine acute inflammation. In these experiments, acute peritonitis was initiated by challenge with the microbial isolate zymosan A, and the actions of five isomers were assessed in vivo. Two compounds (compound V and compound VI) were excluded from matching with PD1 because the physical retention times on LC and GC-MS (Fig. 1 and Table I) and biosynthetic considerations indicated that they were not likely candidates for endogenous human PD1 or isomers produced. It is noteworthy that PD1 (compound II; Fig. 2) at doses as low as 1 ng/mouse gave striking inhibition of PMN infiltration within the exudates. In these experiments, the double dioxygenation product 10S,17S-docosatriene (compound I) was substantially less potent. In this context, the double dioxygenation product was not active at 0.1 ng compared directly to synthetic PD1. At higher doses, 10S,17S-HDHA (compound I) blocked PMN infiltration, but it was less potent than PD1. Compound IV, which is the 10R version of the double dioxygenation products, was essentially equipotent at a 1-ng dose (compound IV compound I), but did not increase potency in a dose-dependent fashion at 10-ng and 100-ng doses (data not shown). The 15-trans isomer of PD1 was, at equal doses of 1 ng/mouse, substantially less potent. Also, a rogue isomer for this series, compound V (Fig. 2) was not likely to be produced in vivo from the 17S-hydroxy precursor because its 10S,17R-diHDHA was essentially without activity in this dose range. Of interest, compound VI was the most potent of these isomers in vivo; however, only trace amounts were noted in human PMN extracts. Hence, a rank order of potency at the 0.1-ng dose of these 10,17-diHDHA isomers was compound VI >> PD1 > 10S,17S-DT (the double dioxygenation product) > the 15-trans-PD1 >> compound V. We also tested the carboxy methyl ester of PD1 vs the native synthetic PD1. Fig. 6B demonstrates the potent dose response of PD1 as it dramatically reduced the infiltration of PMN into the peritoneum. The carbon-1 position carboxy methyl ester was similar in its ability to block in vivo the hallmark of acute inflammation, namely PMN infiltration. The methyl ester of compound VI also proved to be a potent regulator of PMN infiltration.
Can PD1 stop inflammation after its initiation?
Next, we tested whether PD1 or its methyl ester could reduce leukocyte infiltration once inflammation had already been initiated. Results in Fig. 7A indicate that doses as low as 1 ng PD1/mouse diminished infiltration of PMN when administered i.p. following 2 h after challenge with zymosan in vivo. Similar and striking results were obtained with the carboxy methyl ester of PD1, also administered i.p. Hence, once PD1 was given, essentially no further infiltration of PMN into the peritoneum was obtained with essentially >90% blocking of further PMN infiltration to the site. We also addressed whether the anti-inflammatory actions of DHA-derived PD1 and EPA-derived resolvin E1 (RvE1) were synergistic or additive in vivo. RvE1 is derived from EPA and is another omega-3-derived counterregulatory anti-inflammatory lipid mediator that we recently isolated and identified (3, 45). When administered together, RvE1 and PD1 both reduced the infiltration of PMN in vivo during zymosan-induced peritonitis (Fig. 7B). These results indicate that they have a potential additive rather than synergistic anti-inflammatory action when administered together in vivo. We also prepared and tested a chemically more stable form of synthetic PD1, i.e., 15,16-dehydro-PD1, that proved to retain activity in vivo, reducing PMN infiltration, albeit was slightly less potent than the native PD1 (Table II). Of interest, differential counts on light microscopy also revealed that both PD1 and its chemical analog 15,16-dehydro-PD1 reduced PMN infiltration and increased the nonphlogistic recruitment of monocytes and lymphocytes (Table II) while reducing inflammation, a hallmark of resolution (17, 31).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
On identification of 10,17-diHDHA in resolving inflammatory exudates (3, 37) and potent anti-inflammatory actions, it was critical to establish its biosynthesis from DHA. To address this action, we studied isolated human PMN, whole blood, microglial cells, and murine exudates and tissues (1, 3). The isolation and identification of alcohol trapping products indicated the involvement of an epoxide intermediate in the conversion of DHA to 10,17S-diHDHA, a docosatriene containing a characteristic conjugated triene structure involving three of the six double bonds present in this compound. The role of a 16(17)epoxide intermediate generated from the 17S-H(p)DHA precursor was further supported by the identification of two vicinal diols, i.e., 16,17S-dihydroxydocosatrienes present in these LC-MS-MS profiles also generated from DHA (1). The 16(17)epoxy-DHA intermediate could open via nonenzymatic hydrolysis to a racemic mixture, i.e., 16R/S,17S-diHDHA, or to a single 16,17S-vicinol alcohol by the actions of an appropriate epoxide hydrolase in a reaction similar to that demonstrated earlier in the biosynthesis of LXA4 (39, 44, 46). The biosynthesis of PD1 by human cells (compound II) with this stereochemistry from a 16(17)epoxide intermediate would require an enzymatic reaction to move the double bond configuration to set the triene geometry to 11E,13E,15Z and direct the attack of H2O and insertion of its oxygen into the carbon-10 position of PD1 determined in the present experiments to be in the 10R stereochemical configuration.
In addition to PD1 (compound II) in human cell extracts, which carried potent bioactions, an isomer 10S,17S-diHDHA (compound I) was also identified in murine exudates with lesser amounts in isolated human cells (Fig. 1). Compound I was found to be a double dioxygenation product and was also formed from DHA, but in a reaction that required two sequential lipoxygenation steps and oxygen incorporation that was directed at the carbon-10 position derived from molecular oxygen (i.e., 18O2 in an enriched atmosphere in vitro). This reaction producing 10S,17S-diHDHA is markedly different from the proposed enzymatic hydrolysis of the epoxide intermediate in mammalian tissues to produce PD1. The double dioxygenation product formed in vivo is different from PD1 in four key ways: 1) PD1 carbon-10 position alcohol is predominantly in the 10R configuration, whereas the dioxygenation product is mainly in the 10S configuration; 2) the double bond structure of PD1 conjugated triene is in the 11E,13E,15Z configuration, and the 10S,17S-dioxygenation product conjugated triene system is in the 11E,13Z,15E configuration; 3) most importantly, PD1 is more potent than the dioxygenation product (PD1 (compound II) >> 10S,17S-diHDHA (compound I)); and 4) PD1 is generated by isolated human leukocytes and tissues.
Also in support of the stereospecific basis of these DHA-derived products in human and murine systems is the bioaction of the 15-trans-PD1 isomer (compound III), which can arise via workup-induced isomerization of PD1 and possesses little bioactivity in vitro or in vivo within the dose or concentration range (Fig. 5) observed with biogenic or synthetic PD1 (Figs. 6 and 7). Also, compound IV, identified in human leukocytes (Fig. 1) and which differed from compound I at the 10R position and carried the same double bond geometry, was essentially equipotent at a 1-ng dose (Fig. 2). Hence, the biosynthesis of PD1 from DHA, from the results of the present experiments, appears to require stereoselective enzymatic steps to evoke bioactions. This requirement for stereoselective enzymatic reactions is widely appreciated in the biosynthesis of eicosanoids (9, 44). The nature of the PD1 epoxide hydrolase in vivo is therefore of interest, particularly in view of the bioactivity results with compound VI, which was the most potent of the isomers (Fig. 6). Compound VI shares the triene geometry of PD1, differing only in the C10 chirality in the S configuration. However, only trace amounts of this isomer appear to be generated by human cells (vide infra). Thus, the fidelity of the enzyme that produces PD1 from the proposed carbonium cation intermediate (1), in its ability to direct insertion of H2O-derived alcohol at carbon-10 exclusively in the 10R with apparently trace amounts of 10S as in compound VI (Figs. 2 and 6), is an intriguing point for further studies.
Earlier results indicated that DHA, which is not a natural substrate for pt5LO, is converted to 10-HDHA by this enzyme and the double dioxygenation product 10,20-diHDHA (47). In addition, Whelan et al. (48) demonstrated that this plant LO is very versatile with DHA as a substrate and identified multiple monohydroxy-DHA products at carbon positions 4, 7, 8, 11, 13, 14, 16, and 17 to give positional isomers of HDHA; each was an enzymatic product of this flexible enzyme. This regioselectivity also likely reflects the degree of enzyme purity as well as the geographic source of the potato. As with other LOs, we found that potato 5-LOX and soybean 15-LO gave specific diHDHA profiles of products that were dependent on pH, enzyme, and substrate concentrations used in the incubations (3). When the substrates were presented in micellar configuration with the enzymes, hydroperoxy intermediates were converted to epoxides that, on hydrolysis, gave many of the isomers as relatively minor products but were nonetheless in quantities useful for in vitro and in vivo studies (37). These findings were advantageous in the preparation of intermediates (i.e., 7,17-diHDHA, 17S-HDHA, and 17S-H(p)DHA) used in biosynthesis studies and determining the identity and actions of enzymatic products generated in vivo as well as by isolated human cells from DHA (1, 3).
The DHA potato 5-LOX products 10,20-diDHA and 10-HDHA were originally reported by J. Whelan (Ref. 47; doctoral thesis with C. Reddy as advisor). Recently, classic steric analysis of 10S-HDHA and the formation of 10,20-diHDHA and 17-H(p)DHA were reportedly optimized for the plant LOs (49). In the present studies, the double dioxygenation product prepared, matched, and identified in both suspensions of human PMN (Fig. 1A; m/z 261 profile) and in vivo during peritonitis carries its alcohols as expected in the 10S,17S configuration in this diHDHA (compound I). Hence, this natural isomer of PD1 (compound II) formed in vivo from DHA has its 13 position double bond in the cis-configuration (i.e., 13Z) and its 11 in the trans configuration within the conjugated triene portion of the molecule (11E,13Z,15E) and possesses some anti-inflammatory activity in vivo, albeit proved to be much less potent than natural or synthetic PD1 (compound II). The human and murine enzymes(s) involved in the biosynthesis of 10S,17S-diHDHA, the dioxygenation product, have not been determined.
In peritonitis, PD1 significantly reduced PMN infiltration at doses as low as 100 ng/mouse that reached an apparent maximal response at 50% range (1). This level of inhibition of PMN infiltration may be related to PD1’s endogenous anti-inflammatory roles in physiological settings and thus relevant in dampening PMN infiltration in inflammation as a natural mechanism rather than complete inhibition of PMN transmigration, an event that in theory could lead to immune suppression of microbial host defense mechanisms. In the present studies, we confirmed that 7,17-diH(p)DHA (1, 3) and 10,17-diH(p)DHA (49) are both double dioxygenation products (Fig. 4). 18O was incorporated at the carbon-10 position alcohol that originated from enriched atmosphere molecular 18O2 via a LO mechanism (Fig. 4A). The evidence for 18O incorporation at carbon-10 position includes the 2 amu increase in prominent ions in the mass spectrum of 10,17-diHDHA, e.g., m/z 299, 263, 183, 343 (cf Fig. 1, B and C) that was obtained with either 17S-H(p)DHA or 17S-HDHA as substrates. The carbon-10 position alcohol results from lipoxygenation and with native DHA as sequential actions of LO(s) (Fig. 4B), because the chirality at the carbon-10 position is likely in predominantly the S configuration (compound I), the remainder in the 10R configuration as in compound IV.
At the 1 ng/mouse dose, 10S,17S-diHDHA (compound I) did display some activity, but this activity did not increase with higher doses in a statistically significant fashion. Also, this double dioxygenation isomer 10S,17S-diHDHA (compound I) was not active at the 0.1-ng dose compared with PD1. Given the double bond geometry determined in the present study for the conjugated triene unit of PD1 as 11E, 13E, 15Z and carbon-10 position alcohol in the R configuration (Figs. 1 and 2 and Table I), it is likely that, once the 16(17)-epoxide intermediate is produced from 17S-H(p)DHA (1), it is enzymatically subject to hydrolysis. This opens attack of the proposed cation intermediate by water-derived oxygen rather than molecular oxygen in vivo to give the 10R configuration and set the triene double bond configuration to trans,trans,cis geometry at 11E, 13E, 15Z in PD1. This enzymatic mechanism is also supported by identification of epoxide-derived alcohol trapping products in human leukocytes and glial cells and the isolation and identification of two vicinal diols as minor hydrolysis products, namely 16,17-dihydroxydocosatrienes (1). Further studies are warranted to identify the enzyme(s) and establish their role in PD1 biosynthesis as noted above.
In earlier experiments, when administered i.v., 10,17S-docosatriene (PD1, compound II; Fig. 2) was found to be more potent than indomethacin in reducing PMN infiltration in murine peritonitis, i.e., 40% inhibition at 100 ng/mouse (1). Synthetic PD1 in as small a dose as 1 ng/mouse gave 40% inhibition of PMN infiltration that was maintained at the 10-ng and 100-ng doses. Thus, synthetic PD1 (compound II) matched with the natural compound is a potent regulator of PMN infiltration in vivo but does not completely block PMN recruitment, which is consistent with its counterregulatory and autacoid actions and apparently would not compromise host defense via immune suppression of effector cell function. This was also the case with human PMN transmigration, which required 10–100 nM PD1 to reduce PMN transmigration by 35–45% in vitro. Thus, although PD1 stereoselectively reduces PMN transmigration in vitro, given its potent actions in vivo, PD1 (compound II) likely targets additional cell types in vivo to evoke its potent anti-inflammatory actions in vivo (Fig. 7 and Table II). In murine peritonitis, PD1 also regulated both monocyte and lymphocyte traffic to the exudates. Alternatively, these potencies in vivo vs isolated human cells might also reflect species differences. As an inducer of peritonitis, zymosan stimulates the initial formation of many endogenous chemoattractants for PMN, i.e., LTB4, the complement component C5a, chemokines, and cytokines (37). Because PD1 stops PMN recruitment in vivo, it counteracts these several different sets of PMN chemoattractants that regulate trafficking of these cells in vivo (Table II and Ref.37). Given the inherent chemical labilities of PD1, we also prepared and tested a more chemically stable form denoted 15,16-dehydro-PD1 (Table II). Although less potent, chemical stabilization of the conjugated double bonds with an acetylenic form proved useful because the molecule retained activity in vivo (Table II). These results are consistent with the approximately 40% inhibition obtained with a 4,5-acetylenic analog of PD1 (37).
In addition to this LO-initiated route of biosynthesis for PD1, an aspirin-triggered route with a 17R epimer of PD1 (17R series) is generated via acetylated cyclooxygenase and subsequent reactions (3, 32); the complete stereochemistry of this bioactive epimer is in progress. It is of interest to note that compound V, 10S,17R-diHDHA (Fig. 2) was essentially inactive in vivo (Fig. 6A). Whether the many beneficial actions reported for DHA in vitro and with DHA dietary supplementation in humans (21, 24, 25, 26) are linked to the formation and actions of these new families of DHA-derived mediators, protectin and D-series resolvins, is of interest and a timely proposal in view of the importance of uncontrolled inflammation in many widely occurring human diseases. These findings also underscore yet another similarity between the immune and neural systems. Hence, results of the present experiments establish the stereochemistry of PD1 (compound II) and its natural isomers generated by human leukocytes and murine tissues during inflammation. Moreover, they confirm the potent stereoselective anti-inflammatory actions of PD1 and provide new avenues to mark the impact of DHA use/supplementation and its endogenous anti-inflammatory/proresolving actions by monitoring PD1, given its unique physical and biological properties documented in the present report.
Note added in proof.
Subsequent to the submission of this report, a rapid communication appeared in the Journal of Lipid Research that claimed, without results from human cells or tissues, the structural assignment of neuroprotectin D1, and that the biosynthesis proposed by Hong et al. (1) was incorrect. Neuroprotectin D1 is the same as protectin D1, originally discovered by the Serhan group, and the complete stereochemical assignment, established in the present report, proved different from that claimed in the Journal of Lipid Research, which did not report results from mammalian cell biosynthesis or authentic neuroprotectin D1 (see Butovich, I. A. 2005. On the structure and synthesis of neuroprotectin D1, a novel anti-inflammatory compound of the docosahexaenoic acid family. J. Lipid Res. 46: 2311–2314).
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported in part by National Institutes of Health Grants GM38765 and P50-DE016191. 
2 Address correspondence and reprint requests to Dr. Charles N. Serhan, Center for Experimental Therapeutics and Reperfusion Injury, Department of Anesthesiology, Perioperative and Pain Medicine, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02115. E-mail address: cnserhan{at}zeus.bwh.harvard.edu
3 Abbreviations used in this paper: LT, leukotriene; LX, lipoxin; PMN, polymorphonuclear leukocyte; DHA, C22:6, docosahexaenoic acid; EPA, C20:5, eicosapentaenoic acid; NPD1, neuroprotectin D1; LO, lipoxygenase; PD1, protectin D1/neuroprotectin D1; pt5LO, 5-LO from potato; LC, liquid chromatography; MS-MS, tandem mass spectrometry; GC, gas chromatography; 17S-H(p)DHA, 17S-hydro(peroxy)-DHA; NMR, nuclear magnetic resonance; 10S,17S-diHDHA, 10S,17S-dihydroxy-docosa-4Z,7Z,11E,13Z,15E,19Z-hexaenoic acid (the dioxygenation product); RvE1, resolvin E1, 5S,12R,18R-trihydroxy-6Z,8E,10E,14Z,16E-eicosapentaenoic acid.
Received for publication June 10, 2005. Accepted for publication November 4, 2005.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
-3 polyunsaturated fatty acids in dogs. Circulation 99: 2452-2457. 
) represents the mean ± SEM percentage migration of neutrophils compared with vehicle-treated neutrophils for nine separate donors and experiments, each performed in triplicate. The results with 
) are the mean ± SEM obtained for six separate PMN donors, where each point was also in triplicate. *, p < 0.01.
) (compound I; Fig. 2), synthetic PD1 (
) (compound V), compound VI (•), or vehicle alone. *, p < 0.05; n 
), or the rogue isomer (compound VI) methyl ester (
). Peritoneal lavages were obtained at 2 h and leukocytes enumerated. Results are expressed as percentage inhibition compared with mice receiving injections with zymosan A (1 mg) and vehicle alone. *, p < 0.05; n 
