Transcellular biosynthesis of cysteinyl in vivo during mouse peritoneal inflammation

Simona Zarini1, Miguel A. Gijo´ n1, Aaron E. Ransome, Robert C. Murphy2, and Angelo Sala3

Department of Pharmacology, University of Colorado Denver, Mail Stop 8303, 12801 East 17th Avenue, Aurora, CO 80045

Communicated by K. Frank Austen, Brigham and Women’s Hospital, Boston, MA, April 7, 2009 (received for review February 23, 2009)

Leukotrienes (LTs) are lipid mediators of inflammation formed by B4 (LTB4). LTC4 can be metabolized by sequential enzymatic oxidation of . One intriguing aspect of proteolytic hydrolysis to (LTD4) and leukotriene LT production is transcellular biosynthesis: cells expressing 5- E4 (LTE4); these 3 compounds are collectively termed cysteinyl (5LO) form LTA4 and transfer it to cells expressing LTs (cysLTs). The final LT products derived from LTA4 play LTA4 hydrolase (LTA4H) or LTC4 synthase (LTC4S) to produce LTB4 or important roles in inflammation, such as inducing neutrophil LTC4. This process has been demonstrated in vivo for LTB4, but not chemotaxis (LTB4) or causing edema and smooth muscle con- for cysteinyl LTs (cysLTs). We examined transcellular cysLT synthe- traction (LTC4, LTD4) (1, 2). The effects of these are sis during zymosan-induced peritonitis, using bone marrow trans- mediated by G-protein coupled receptors. Two have been iden- plants with transgenic mice deficient in key of LT syn- tified for LTB4 (BLT1 and BLT2) (13, 14), and 3 for cysLTs thesis and analyzing all eicosanoids by liquid chromatography/ (cysLT1, cysLT2, and GPR17) (15–17). Although the protein has tandem mass spectrometry. WT mice time-dependently produced not been identified, a novel showing specificity for ؊/؊ LTB4 and cysLTs (LTC4, LTD4, and LTE4). 5LO mice were incapable LTE4 has been recently described (18). CysLT1 receptor antag- of producing LTs. WT bone marrow cells restored this biosynthetic onists are currently used to help control symptoms, and ability, but 5LO؊/؊ bone marrow did not rescue LT synthesis in drugs that interfere with LT biosynthesis or action are candidates irradiated WT mice, demonstrating that bone marrow-derived cells for preventing cardiovascular disease (19, 20). are the ultimate source of all LTs in this model. Total levels of LT production is regulated at different levels (8), including ؊/؊ 5LO-derived products were comparable in LTA4H and WT mice, methylation of the 5LO promoter (21) and posttranslational ؊/؊ IMMUNOLOGY but were reduced in LTC4S animals. No differences in prosta- control of 5LO activity through phosphorylation, interaction glandin production were observed between these transgenic or with FLAP (11, 12), and translocation induced by calcium or free ؊/؊ ؊/؊ chimeric mice. Bone marrow cells from LTA4H or LTC4S mice AA (22). Additionally, experimental evidence suggests that cells ؊/؊ injected into 5LO mice restored the ability to synthesize LTB4 cooperate in the synthesis of LTB4 and LTC4. For example, and cysLTs, providing unequivocal evidence of efficient transcel- human , which express no 5LO, can interact with lular biosynthesis of cysLTs. These results highlight the potential activated neutrophils, which do not express LTC4S, to generate relevance of transcellular exchange of LTA4 for the synthesis of LTs LTC4 (23). Many other instances of transcellular LT biosynthesis mediating biological activities during inflammatory events in vivo. have been described (24). The chemical half-life of LTA4 in buffer is less than 5 s (25), which implies some protective chimeric mice ͉ ͉ mass spectrometry ͉ ͉ mechanism of the epoxide ring during transfer from the donor 5-lipoxygenase cell to the acceptor cell. Mice deficient in enzymes responsible for LT synthesis are eukotrienes (LTs) are biologically active lipid mediators that valuable tools to test the extent of transcellular formation of LTs. Koller and colleagues illustrated this value by transplanting bone Lplay important roles in inflammation and are involved in Ϫ/Ϫ Ϫ/Ϫ pathological states with an inflammatory component, such as marrow from LTA4H mice into lethally irradiated 5LO asthma, cardiovascular disease, or cancer (1–5). They derive mice (26). In a peritoneal inflammation model, LTB4 was found from arachidonic acid (AA) through the action of 5-lipoxygen- to be synthesized, proving that transcellular biosynthesis did ase (5LO) (6), an expressed in a limited number of cells, occur in vivo. Production of LTB4 in these chimeric animals was significantly lower than in WT controls or 5LOϪ/Ϫ mice rescued including neutrophils, eosinophils, monocytes, macrophages, with WT bone marrow. This raised questions about the extent of mast cells, and basophils (7, 8). Considerable work has led to our transcellular biosynthesis, whether this was a minor pathway for current understanding of events regulating the formation of in vivo LT generation, or whether LTA was converted into other these mediators. AA is released from phospholipids by cytosolic 4 products besides LTB . phospholipase A ␣ (cPLA ␣) after this enzyme translocates 4 2 2 In the present study, our main aim was to find evidence of the from cytosol to perinuclear membranes of cells, usually following possible transcellular biosynthesis of cysLTs using LTC SϪ/Ϫ an increase in calcium ion concentration (9). 5LO, a soluble 4 bone marrow cells, while identifying and quantifying all of the protein in the cytosol or nucleoplasm of resting cells, also major COX- and 5LO-derived eicosanoids. translocates to perinuclear membranes when calcium levels increase (10). 5LO oxidizes free AA to 5-hydroperoxyeicosatet- Results raenoic acid. This intermediate can be reduced by peroxidases to Generation of Eicosanoids in Zymosan-Induced Peritonitis. Injection 5-hydroxyeicosatetraenoic acid (5-HETE) or dehydrated, in a of zymosan into the peritoneal cavity of mice resulted in second 5LO-catalyzed reaction, to (LTA4). 5LO action is facilitated by 5LO-activating protein (FLAP), an inte- gral nuclear membrane protein, and the assembly of a multien- Author contributions: S.Z., M.A.G., R.C.M., and A.S. designed research; S.Z., M.A.G., A.E.R., zymatic machine to make LTs (10–12). Thus, LTA4 is formed and A.S. performed research; S.Z., M.A.G., R.C.M., and A.S. analyzed data; and S.Z., M.A.G., near the , and this chemically reactive molecule R.C.M., and A.S. wrote the paper. must find its way either to synthase (LTC4S), The authors declare no conflict of interest. another nuclear membrane protein that conjugates LTA4 with 1S.Z. and M.A.G. contributed equally to this work. to form leukotriene C4 (LTC4), or to cytosolic LTA4 2To whom correspondence should be addressed. E-mail: [email protected]. hydrolase (LTA4H), which stereospecifically opens the epoxide 3Present address: Department of Pharmacological Sciences, University of Milan, Via Balza- ring and controls the addition of water at carbon-12 to form retti 9, 20133 Milan, Italy.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903851106 PNAS Early Edition ͉ 1of6 Downloaded by guest on September 27, 2021 LTE4 5 (m/z 438→333)

6 trans Δ - -LTB4s (m/z 335→195) LTB4 5-HETE 4 m/z ( 335→195) (m/z 319→115)

) LTD4 -3 (m/z 495→177)

3 LTC4 (m/z 624→272)

PGE2 2 (m/z 351→271) 5,6-diHETEs (m/z 335→115) TXB2 (m/z 369→169) Intensity (cps x 10 Intensity 1

0 5 10 15 20 25 Time (min)

Fig. 1. LC/MS/MS profile of eicosanoids in peritoneal lavage. Mice were treated with 1-mg zymosan for 2 h, and eicosanoids analyzed by LC/MS/MS, monitoring the specific m/z transitions indicated. Retention times of deuterated internal standards (described in the text) coincided with their cognate metabolites.

production of eicosanoids, as revealed by liquid chromatogra- with calcium ionophore A23187 (Table 1). In unchallenged mice, phy/tandem mass spectrometry (LC/MS/MS) (Fig. 1). After 2 h, peritoneal cells produced mainly LTC4, consistent with the eicosanoids derived from the 5LO and COX pathways of AA presence of macrophages and mast cells. Upon zymosan treat- oxidation were detected. The most abundant metabolites in ment, there was a time-dependent decrease of cysLT production peritoneal lavage were LTE4, LTB4, and 5-HETE. Also present and an increase in LTB4 production, likely because of macro- 6 were LTC4, LTD4, ⌬ -trans-LTB4 isomers, 5,6-dihydroxyeicosa- phages being replaced by neutrophils. tetraenoic acid (diHETE) isomers, and small amounts of COX- derived B2 (TXB2) and E2 (PGE2). Production of Leukotrienes in Transgenic Mice Deficient in Leukotri- Levels of 5,6-diHETEs were consistently higher (2–8 times) than ene Biosynthetic Enzymes. Production of eicosanoids by WT and 6 6 levels of ⌬ -trans-LTB4s. In subsequent experiments, ⌬ -trans- transgenic mice was investigated (Fig. 3). As expected,2hafter Ϫ/Ϫ LTB4s and 5,6-diHETEs are reported together as nonenzymatic zymosan injection the peritoneal lavage of 5LO mice showed products of LTA4. 6-keto-PGF1␣ and PGF2␣ were barely de- no LTB4 and barely detectable amounts of cysLTs. The levels of tected above background (10–20% of the PGE2 signal). PGD2, COX-derived TXB2 and PGE2 were similar to WT mice. No Ϫ/Ϫ PGJ2, PGA2, 20-OH-LTB4, and 20-COOH-LTB4 were below the other COX-derived metabolite was increased in 5LO ani- limit of detection. 12-HETE and 15-HETE were also found, but mals, indicating that no shunting of AA substrate between the 2 their levels were similar in control mice injected with PBS. None oxygenases occurred in this model. These animals showed of the other eicosanoids was detected in mice injected with PBS decreased MPO activity (indicative of neutrophil infiltration) or not treated at all (Fig. 2, 0 h time point). and protein concentration (indicative of plasma extravasation) in Zymosan-induced LT release was time-dependent. LTB4 in- the peritoneal lavage (Table 2), in line with sharply reduced LT Ϫ/Ϫ creased at 1 h and peaked at 2 h (see Fig. 2, Top). 5-HETE and levels. In LTA4H mice, no LTB4 was produced, but levels of nonenzymatic LTA4 metabolites followed similar time courses. cysLTs, 5-HETE and nonenzymatic products of LTA4 were In contrast, cysLTs peaked at 1 h and were decreased at 2 h, with slightly increased compared to WT animals. Total production of LTE4 remaining elevated at 3 h (see Fig. 2, Bottom). Subsequent LTA4 (the sum of all LTA4-derived eicosanoids) was comparable Ϫ/Ϫ experiments were carried out at2htoassess the effects of between WT and LTA4H mice (see Fig. 3, Inset). In the case Ϫ/Ϫ genetic manipulation in both LTB4 and LTE4 levels. of LTC4S mice, normal levels of LTB4, nonenzymatic LTA4 The initial release of eicosanoids was likely the result of products and 5-HETE were measured. Virtually no cysLTs were zymosan phagocytosis by resident peritoneal macrophages or detected, and the total levels of LTA4 produced were signifi- mast cells, the predominant 5LO-expressing cells recovered cantly lower than in WT mice. The reasons for this are not known from peritoneal cavity of unchallenged mice (27). LTB4 gener- at this time, but they could not be explained by differences in cell ated in the first hour of stimulation possibly helped recruit populations in the peritoneal cavity. In additional experiments neutrophils to the peritoneal cavity, which would explain LTB4 with unchallenged mice, both the total number of peritoneal cells release at later time points. Indeed, zymosan increased MPO recovered and the relative content of monocytes/macrophages Ϫ/Ϫ activity in the peritoneal lavage (see Fig. 2). At 3 h, MPO was and mast cells were very similar between WT and LTC4S maximal, but LTB4 synthesis was lower than at 2 h, probably mice. because of zymosan depletion. It has been shown that a second injection of zymosan6hafterstimulation results in a second, Generation of Leukotrienes in Chimeric Mice. A major strategy of more pronounced, burst of LTB4 production in the peritoneal this study was to restrict the presence of 5LO to cells of bone cavity of mice, consistent with increased numbers of neutrophils marrow origin, as opposed to cells of nonhematopoietic origin (28). Further evidence linking cell population changes and LT present in the peritoneal cavity. This was accomplished by production was provided by stimulating peritoneal lavage cells lethally irradiating 5LOϪ/Ϫ mice and rescuing the animals with

2of6 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903851106 Zarini et al. Downloaded by guest on September 27, 2021 60 10 * WT LTB4 5LO-/- NonEnz 50 LTA H-/- 8 5HETE 4 LTC4S-/-

Total LTA4-Derived Products 40 * 6 * 30 4 pmol/ml lavage pmol/ml

* pmol/ml lavage 20 2 10 0 120 LTC 0 4 * LTB NonEnz 5HETE LTC LTD LTE TXB PGE LTD4 4 4 4 4 2 2 100 LTE4 Fig. 3. Production of eicosanoids in peritoneal lavage of transgenic mice deficient in LT synthesis enzymes. WT or transgenic mice with the indicated 80 genotypes were treated with 1-mg zymosan for 2 h, and the indicated eico- sanoids were analyzed by LC/MS/MS and quantitated by isotopic dilution of 60 the corresponding internal standards. Data in the inset represent the sum of 6 * all LTA4-derived metabolites (⌬ -trans-LTB4s and 5,6-diHETEs, LTB4, and cys- * LTs). Results are average Ϯ SEM of at least 3 experiments. , P Ͻ 0.05 vs. WT. 40 * pmol/ml lavage lavage pmol/ml

20 through transcellular of LTA4 by stromal cells within the bone marrow cell preparation. In stimulated bone marrow IMMUNOLOGY Ϫ/Ϫ Ϫ/Ϫ 0 cells from LTC4S 3 5LO , LTC4 was undetectable, MPO 0.03±0.00 0.04±0.01 0.24±0.05 0.97±0.11 Ϫ/Ϫ whereas WT or LTA4H cells produced measurable levels of 0 1 2 3 this cysLT, consistent with previous results (29). Ϫ/Ϫ Time (h) When irradiated 5LO mice were rescued with WT cells, zymosan triggered abundant production of both LTB4 and LTE4 Fig. 2. Time-course of production of 5LO-derived metabolites and myelo- (Fig. 4). Nonenzymatic hydrolysis products of LTA4 were also peroxidase (MPO) activity. Mice were treated with 1-mg zymosan for the times detected, as were 5-HETE, LTC4, and LTD4. Phagocytic stim- indicated, and eicosanoids were analyzed by LC/MS/MS and quantitated by ulation clearly activated the 5LO pathway, resulting in formation isotopic dilution of the corresponding internal standards. NonEnz: sum of ⌬6 Ͻ of LTA4. The amounts of LTB4 and LTE4 were not significantly 5,6-diHETEs and -trans-LTB4s. *, P 0.05 vs. 0 h, for LTB4 and LTE4. Also shown is the MPO activity (Abs 570 nm) in peritoneal lavage at the same time different from WT animals that had not undergone irradiation points. Results are average Ϯ SEM of at least 4 experiments. or transplantation. Neutrophil infiltration (as indicated by MPO activity) and protein concentration were also similar to those observed in nonirradiated WT animals (see Table 2). bone marrow cells from mice that did express 5LO, either WT, When peritonitis was induced in chimeric animals with a Ϫ Ϫ Ϫ Ϫ / / Ϫ/Ϫ Ϫ/Ϫ LTA4H ,orLTC4S . The success of the transplants was 5LO background and bone marrow from LTA4H animals, confirmed by the ability of the regenerated bone marrow cells to significant amounts of LTB4 were observed (see Fig. 4). Un- synthesize eicosanoids upon stimulation with A23187 (Table 3). equivocally, this could only be produced in these In 42 out of 43 chimeric animals, the eicosanoid profile showed animals through transcellular metabolism, because no cell could that, after the 8-week period of recovery, the bone marrow express both 5LO and LTA4H (Fig. 5). However, the levels of Ϫ/Ϫ population exhibited the expected phenotype of the donor rather LTB4 were not as high as those in WT mice or WT 3 5LO Ϫ/Ϫ than the recipient mouse, validating the irradiation and implan- chimeras. This suggested the possibility that LTA4H 3 Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ tation protocols. Bone marrow cells from LTA4H 3 5LO 5LO chimeras may exhibit decreased 5LO activity, but non- mice released abundant 5LO-derived products but, compared to enzymatic LTA4 products, 5-HETE, and cysLTs, were measured Ϫ/Ϫ WT or LTC4S cells, very small amounts of LTB4, possibly Table 2. Protein concentration and MPO activity in Table 1. Eicosanoid production by peritoneal lavage cells peritoneal lavage 0h 1h 2h 3h Protein (mg/ml) MPO (Abs 570 nm) Ϯ Ϯ LTB4 1.17 Ϯ 0.27 0.50 Ϯ 0.03 2.38 Ϯ 0.33 7.80 Ϯ 1.67 WT 2.52 0.07 0.23 0.05 6 Ϫ/Ϫ Ϯ Ϯ ⌬ t-LTB4s 3.29 Ϯ 0.74 0.43 Ϯ 0.02 1.21 Ϯ 0.18 3.77 Ϯ 0.77 5LO 1.38 0.07* 0.03 0.01 Ϫ/Ϫ 5,6-diHETEs 2.20 Ϯ 0.47 0.34 Ϯ 0.04 0.77 Ϯ 0.11 1.88 Ϯ 0.13 LTA4H 3.56 Ϯ 0.38* 0.11 Ϯ 0.01 Ϫ/Ϫ 5-HETE 6.59 Ϯ 1.17 0.88 Ϯ 0.06 3.15 Ϯ 0.47 8.18 Ϯ 0.57 LTC4S 2.16 Ϯ 0.08* 0.37 Ϯ 0.11 3 Ϫ/Ϫ Ϯ Ϯ LTC4 19.11 Ϯ 3.51 0.92 Ϯ 0.16 0.54 Ϯ 0.12 0.48 Ϯ 0.27 WT 5LO 2.61 0.17 0.40 0.13 Ϫ/Ϫ 3 Ϯ Ϯ TXB2 2.20 Ϯ 0.44 0.14 Ϯ 0.02 0.36 Ϯ 0.06 1.31 Ϯ 0.11 5LO WT 1.17 0.06* 0.12 0.05 Ϫ/Ϫ 3 Ϫ/Ϫ Ϯ Ϯ PGE2 2.55 Ϯ 0.58 0.19 Ϯ 0.04 1.47 Ϯ 0.27 3.99 Ϯ 0.55 LTA4H 5LO 2.41 0.13 0.24 0.12 Ϫ/Ϫ Ϫ/Ϫ LTC4S 3 5LO 3.03 Ϯ 0.40 0.41 Ϯ 0.19 WT mice were treated with 1-mg zymosan for the times indicated, and peritoneal lavage cells (106 in 1 ml) were stimulated with 0.5 ␮M A23187. All Mice were treated with 1-mg zymosan for 2 h. Protein concentration and eicosanoid products were measured by LC/MS/MS. Results (ng/106 cells) are MPO activity were determined in the peritoneal lavage. Results are expressed average Ϯ SEM of at least 5 experiments. as average Ϯ SEM of at least 3 experiments. *, P Ͻ 0.05 vs. WT.

Zarini et al. PNAS Early Edition ͉ 3of6 Downloaded by guest on September 27, 2021 Table 3. Eicosanoid production by bone marrow cells in chimeric mice

Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ WT 3 5LO 5LO 3 WT LTA4H 3 5LO LTC4S 3 5LO

LTB4 7.39 Ϯ 1.20 n.d. 0.65 Ϯ 0.24 5.90 Ϯ 2.80 6 ⌬ t-LTB4s 2.03 Ϯ 0.30 n.d. 4.46 Ϯ 0.67 1.47 Ϯ 0.62 5,6-diHETEs 1.49 Ϯ 0.18 n.d. 1.40 Ϯ 0.29 1.05 Ϯ 0.76 5-HETE 7.33 Ϯ 1.19 n.d. 6.62 Ϯ 1.32 5.10 Ϯ 2.59

LTC4 0.20 Ϯ 0.04 n.d. 0.39 Ϯ 0.05 n.d. TXB2 1.79 Ϯ 0.22 3.10 Ϯ 1.39 1.40 Ϯ 0.17 1.76 Ϯ 0.25 PGE2 9.77 Ϯ 1.25 10.05 Ϯ 0.81 9.71 Ϯ 2.50 9.28 Ϯ 0.75

Bone marrow cells (106 in 1 ml) from chimeric mice (donor 3 acceptor) were stimulated with 0.5 ␮M A23187. All eicosanoid products were measured by quantitative LC/MS/MS. Results (ng/106 cells) are average Ϯ SEM of at least 6 experiments. n.d., not detected.

at levels comparable to WT animals (see Fig. 4, Inset), indicating cysLTs. It has been reported that macrophages and mast cells are that absence of LTA4H did not affect other components of LT relatively radioresistant (30, 31). Compared to the robust res- synthesis. In these chimeric mice, MPO levels were higher than toration of LT synthesis in the chimeric animals described above, Ϫ/Ϫ Ϫ/Ϫ in LTA4H animals (see Table 2), consistent with LTB4 being these experiments with 5LO bone marrow provide solid present in the peritoneal lavage. support for the hypothesis that the overwhelming majority of Ϫ/Ϫ Ϫ/Ϫ For 5LO animals transplanted with LTC4S cells, the LTA4 generated during zymosan-induced peritonitis in the most striking observation was the significant production of mouse is derived from bone marrow hematopoietic cells. LTE4, only possible through transcellular exchange of LTA4 (see Fig. 4). In these animals, no single cell expressed both 5LO and Discussion LTC4S (see Fig. 5). To our knowledge, this report of transcellular Eicosanoids, such as PGs and LTs, are potent lipid mediators synthesis of cysLTs in an in vivo model of inflammation is involved in inflammatory responses. It is not surprising that their unique. Moreover, this process was very efficient, as total synthesis is tightly regulated through a variety of mechanisms, production of cysLTs (almost 14 pmol of LTC4, LTD4, and some of which are not well understood, including transcellular LTE4) was more than 50% of the total LTA4 production (around biosynthesis. In this study, we revisited the zymosan-induced 25 pmol of cysLTs, LTB4, and nonenzymatic LTA4 products). peritonitis model used to demonstrate transcellular LTB4 syn- Ϫ/Ϫ Protein concentration in the peritoneal lavage of LTC4S 3 thesis (26), using additional transgenic mice and comprehensive Ϫ/Ϫ Ϫ/Ϫ 5LO mice was higher than LTC4 animals, and similar to measurement of eicosanoids to help deepen our understanding WT, correlating with the presence of cysLTs. This suggested that of this phenomenon. cysLT transcellular biosynthesis may be sufficient to induce When acute peritoneal inflammation was induced in mice by plasma extravasation in this experimental model. zymosan, the major LTs observed after 1 h were cysLTs, In contrast with the production of LTB4 and cysLTs in the predominantly LTE4. We did not analyze lavages at shorter peritoneal cavity of WT animals, chimeric animals of a WT times, but it has been shown that LTC4 is the main LT after 15 background, lethally irradiated and rescued with 5LOϪ/Ϫ cells, min of i.p. injection of either opsonized (32) or nonopsonized were found to generate no LTB4 and very low levels of cysLTs. (33) zymosan, with LTE4 becoming more abundant after 1 h. There was a small but reproducible amount of LTE4, suggesting that a small number of cells in the peritoneal cavity of WT animals survived irradiation and remained able to generate Bone marrow-derived cells

LTA H-/- LTC S-/- AA 4 AA 4 70 WT 5LO-/- → * 5LO 5LO 5LO-/- → WT LTA H-/- 5LO-/- cys-LTs 60 4 → LTA4 LTA4 LTB4 LTC4S LTA4H LTC4S-/- → 5LO-/-

50 Total LTA4-Derived Products

40 LTB4 LTA H 30 4 LTA4

pmol/ml lavage * LTC4S 20 cys-LTs 5LO-/- 10 Peritoneal parenchyma cells

0 Fig. 5. Experimental model of transcellular leukotriene biosynthesis. 5LOϪ/Ϫ

LTB4 NonEnz 5HETE LTC4 LTD4 LTE4 TXB2 PGE2 mice were lethally irradiated, then rescued with bone marrow cells from Ϫ/Ϫ Ϫ/Ϫ LTA4H or LTC4S mice. Upon zymosan stimulation, cells of a bone marrow Ϫ/Ϫ Fig. 4. Production of eicosanoids in peritoneal lavage of chimeric mice. Chi- origin could synthesize either cysLTs (if donors were LTA4H )orLTB4 (if Ϫ/Ϫ meric mice (bone marrow donor 3 irradiated acceptor) were treated with 1-mg donors were LTC4S ), but no single cell could synthesize both. Parenchymal zymosan for 2 h, and the indicated eicosanoids were analyzed by LC/MS/MS and cells could not synthesize either, as they did not express 5LO. To produce both quantitated by isotopic dilution of the corresponding internal standards. The types of LTs, LTA4 needed to be exported from bone marrow-derived cells and LTB4 section is expanded for clarity. The inset shows the sum of all LTA4-derived imported into parenchymal cells, then converted to LTB4 and cysLTs. LTA4H 6 metabolites (⌬ -trans-LTB4 isomers and 5,6-diHETE isomers, LTB4 and cysLTs). and LTC4S were both expressed in the parenchymal cell population, although Results are average Ϯ SEM of at least 3 experiments. *, P Ͻ 0.05. not necessarily within the same cell.

4of6 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903851106 Zarini et al. Downloaded by guest on September 27, 2021 After 2 h, LTB4 was also abundant in peritoneal lavage, most Transcellular synthesis of LTB4 and cysLTs represents only a likely because of neutrophil infiltration. Assessment of the fraction of the LTs produced in WT mice (Ϸ20–25%). These peritoneal lavage cell ability to produce eicosanoids upon stim- results are remarkably similar to those published by Koller and ulation and MPO activity provided evidence of a change in cell colleagues (26), where transcellular synthesis of LTB4 accounted population in agreement with previous reports on this model (27, for a significant portion of plasma extravasation and edema 28, 34, 35). All these cells have a hematopoietic origin, and while during AA-induced cutaneous inflammation. Our data on MPO other cells in the peritoneal cavity might possibly express 5LO, activity and protein concentration suggest that LTB4 and cysLTs data from irradiated WT mice rescued with 5LOϪ/Ϫ cells produced transcellularly in this peritonitis model contribute demonstrate that none of the LTA4 in this peritonitis model substantially to neutrophil recruitment and plasma extravasa- originates in cells not derived from bone marrow. Likely sources tion, respectively, supporting the potential biological relevance of LTA4 are peritoneal monocytes/macrophages, mast cells, or of transcellular LT synthesis. influxing neutrophils. Although mast cells are known to release In summary, this study provides evidence for the occurrence LTs in response to zymosan (36), previous reports suggest that of transcellular biosynthesis of cysLTs in vivo. Given the effi- they are involved in LTB4 rather than cysLT production (28, 33). ciency of transfer of the reactive intermediate LTA4 to generate The results presented here also show that parenchyma cells of cysLTs, it appears that transcellular synthesis could be a sizable nonhematopoietic origin express downstream enzymes in the LT component of cysLT production during inflammation and con- synthesis pathway, such as LTA4H and LTC4S. tribute to the overall inflammatory response. The separation of The mechanism by which LTA4 is transported between cells is critical enzymes in different cells probably constitutes an addi- poorly understood. This chemically reactive epoxide must be tional regulatory mechanism controlling production of these stabilized in a manner that shields it from the aqueous environ- important mediators, and as such it provides a potentially unique ment, because water rapidly hydrolyzes LTA4 in a nonenzymatic target for pharmacological intervention. 6 process leading to the production of ⌬ -trans-LTB4s (6-trans- LTB4 and 6-trans-12-epi-LTB4) and, to a lesser extent, 5,6- Materials and Methods diHETEs (37). In this model, however, higher levels of 5,6- Chemicals. All chemicals and solvents were purchased from Fisher Scientific. diHETEs were observed, indicating preferential attack of water Zymosan, BSA, hexadecyltrimethylammonium bromide (HTAB), and 3,3Ј,5,5Ј- at carbon-6 of LTA4 rather than at carbon-12, the thermody- tetramethylbenzidine (TMB) were from Sigma-Aldrich. Stable isotope-labeled Ն namically most stable position for water to attack the delocalized internal standards [d4]LTB4 ( 97 atom %D), [d5]LTC4 (97%D), [d8]5-HETE Ն Ն Ն carbocation after the epoxide ring is opened. This suggests that ( 98%D), [d4]TXB2 ( 98%D), [d4]PGE2 ( 99%D), as well as synthetic LTs used IMMUNOLOGY in stable isotope dilution curves, were purchased from Cayman Chemical. some protein is binding LTA4 through a hydrophobic pocket,

which protects carbon-12 to a larger extent than carbon-6. Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Animals. 5LO and LTA4H mice (129S6 background) were a generous gift When irradiated 5LO mice were rescued with LTA4H Ϫ/Ϫ from B.H. Koller (University of North Carolina) (43, 34). LTC4S mice (C57BL bone marrow, LTB4 was produced transcellularly in quantities background) were a generous gift from K.F. Austen, B.K. Lam, and Y. Kanaoka significantly lower than in WT animals, in agreement with the (Harvard Medical School) (44). 129S6 mice were from Taconic. C57BL mice were report from Koller and colleagues (26). A possible explanation from The Jackson Laboratory. All animal experiments were approved by the would be that most of the LTA4 was hydrolyzed nonenzymati- Institutional Animal Care and Use Committee at the University of Colorado cally and did not reach a cell that had LTA4H. Alternatively, Denver. chimeras may have generated less LTA4 than WT mice. In this Ϫ/Ϫ Ϫ/Ϫ study, total production of LTs in LTA4H 3 5LO chimeric Mouse Peritoneal Inflammation Model. Zymosan was boiled for 5 min in HBSS, mice was found to be quite comparable to WT mice, suggesting resuspended at 1 mg/ml, and injected (1 ml) into the peritoneal cavity of mice. that there was not a major change in the cell population or the After 1, 2, or 3 h, mice were killed and the peritoneal cavity washed with 5 ml activity of 5LO. The loss of LTB was balanced by an increase HBSS. One milliliter of the lavage fluid was added to 1 ml of methanol for 4 eicosanoid measurement by LC/MS/MS. Separate aliquots (200 ␮l) were used in cysLTs and nonenzymatic LTA4 products. In contrast, Ϫ/Ϫ for protein measurement (bicinchoninic acid assay, Pierce) and MPO assay. For LTC4S mice, and chimeras generated with their bone mar- the assay, cells were resuspended in 50-mM phosphate buffer, pH 6, 0.1% row, showed reduced levels of LTA4 and other metabolites gelatin, 0.5% HTAB, and incubated with TMB Liquid Substrate for 20 min, derived from 5LO. The reasons for this surprising finding may before monitoring absorbance at 570 nm. include decreased expression of 5LO or FLAP, or some un- Cells in the remaining lavage were pelleted, resuspended in HBSS, and known mechanism regulating 5LO activity in peritoneal cells. It stimulated at a concentration of 106/ml (1 ml) with 0.5 ␮M calcium ionophore is interesting to note that physical interaction between 5LO, A23187 at 37 °C. After 10 min, 1 ml of methanol was added for extraction and LC/MS/MS analysis. FLAP, and LTC4S has been reported (12), and the absence of one of these enzymes during cell differentiation may affect the stability of the others. As an additional conclusion, the fact that Eicosanoid Analysis by LC/MS/MS. After addition of stable isotope-labeled LTC SϪ/Ϫ bone marrow cells were unable to synthesize LTC internal standards, eicosanoids in peritoneal lavages or stimulated cells were 4 4 extracted and analyzed essentially as previously described (29), with some upon A23187 stimulation, indicates that LTC4S was the pre- modifications: [d5]LTC4 (5 ng) was used as an internal standard for cysLTs; solid dominant enzyme able to convert LTA4 into LTC4 in this model, phase cartridges were Strata-X, 33 ␮m Polymeric Reversed Phase (Phenome- suggesting that microsomal GSTs (38, 39) are not involved. nex); chromatography was performed on an Ascentis 150 ϫ 2 mm, 5 ␮m Transcellular synthesis of cysLTs has been shown in heterol- column (Supelco), at 200 ␮l/min with a linear gradient from 45% solvent B to ogous perfusion models with human neutrophils and rabbit 75% in 12 min, 75% to 98% in 2 min, 11 min hold and re-equilibration for 10 hearts or guinea pig brains (40, 41). This report provides min; and m/z transitions were included for additional eicosanoids. Transitions 3 ⌬6 3 unequivocal evidence of transcellular cysLT production in an in monitored were: 335 195, LTB4 and -trans-LTB4s; 351 195, 20-OH-LTB4; 3 3 3 3 vivo model of inflammation in a single organism. A very 365 195, 20-COOH-LTB4; 335 115, 5,6-diHETEs; 624 272, LTC4; 495 177, 3 3 3 3 Ϫ/Ϫ 3 LTD4; 438 333, LTE4; 319 115, 5-HETE; 319 179, 12-HETE; 319 219, 15- surprising result was that production of cysLTs in LTC4S 3 3 3 5LOϪ/Ϫ chimeric mice accounted for more than half the total HETE; 351 271, PGE2 and PGD2; 353 309, PGF2␣; 333 189, PGJ2 and PGA2; 3693163, 6-keto-PGF1␣; 3693169, TXB2; 6293272, [d5]LTC4; 3273116, [d8]5- LTA4 produced. Because the only way in which cysLTs could be HETE; 3393197, [d ]LTB ; 3733173, [d ]TXB ; and 3553275, [d ]PGE . generated was transcellular biosynthesis, it is reasonable to 4 4 4 2 4 2 conclude that transcellular production of cysLTs is a very Generation of Chimeric Animals. After fasting for 16 h, mice were lethally efficient process in this model. This is consistent with the fact irradiated (RS-2000 X-ray Biological Irradiator, Rad Source Technologies) in a that more than 50% of LTA4 produced by human neutrophils is ventilated container, using 2 cycles of 300 rads with a 3-h interval. Bone marrow exported from the cell (42). cells were isolated from femurs and tibias of donor mice as previously described

Zarini et al. PNAS Early Edition ͉ 5of6 Downloaded by guest on September 27, 2021 (29). The cell suspension was filtered through a 40-␮m nylon cell strainer and peritoneal lavage (or ng for stimulated cells) and are reported as the aver- centrifuged at 160 ϫ g for 10 min. Cells were resuspended in PBS containing 0.5% age Ϯ SEM, with statistical significance considered as P Ͻ 0.05. Unpaired BSA at a density of 107/ml. Immediately after irradiation, recipient animals were Student’s t test was performed using GraphPad InStat 3. injected through the tail vein with 200 ␮l of bone marrow cell suspension using a syringe with a 29-gauge needle. Mice were then fed and housed normally for ACKNOWLEDGMENTS. We thank Dr. Beverly H. Koller (University of North 8 weeks. To verify the irradiation/transplantation protocol, bone marrow cells Carolina) and Drs. K. Frank Austen, Bing K. Lam, and Yoshihide Kanaoka from each transplanted animal were harvested after collection of the peritoneal (Harvard Medical School) for kindly providing transgenic mice. Dr. Koller and 6 ␮ lavage. Cells (10 in 1 ml) were stimulated with 0.5 M A23187, and eicosanoids Dr. Peter Henson (National Jewish Health) contributed valuable discussion measured as described above. regarding the manuscript. This work was supported by National Institutes of Health Grant HL025785. A.S. received partial support from the European Data Analysis. Eicosanoids were quantitated with the help of MultiQuant Community (LSHM-CT-2004–005033) and was the recipient of a William Ful- software (Applied Biosystems/MDS SCIEX). Data are expressed as pmol/ml bright Research Scholarship.

1. Peters-Golden M, Henderson WR, Jr (2007) Leukotrienes. N Engl J Med 18:1841–1854. 26. Fabre JE, et al. (2002) Transcellular biosynthesis contributes to the production of 2. Nicosia S, Capra V, Rovati GE (2001) Leukotrienes as mediators of asthma. Pulm leukotrienes during inflammatory responses in vivo. J Clin Invest 109:1373–1380. Pharmacol Ther 14:3–19. 27. Mullaly SC, Kubes P (2007) Mast cell-expressed complement receptor, not TLR2, is the 3. De Caterina R, Zampolli A (2004) From asthma to atherosclerosis–5-lipoxygenase, main detector of zymosan in peritonitis. Eur J Immunol 37:224–234. leukotrienes, and inflammation. N Engl J Med 350:4–7. 28. Rao TS, Currie JL, Shaffer AF, Isakson PC (1994) In vivo characterization of zymosan- 4. Zhao L, Funk CD (2004) Lipoxygenase pathways in atherogenesis. Trends Cardiovasc induced mouse peritoneal inflammation. J Pharmacol Exp Ther 269:917–925. Med 14:191–195. 29. Gijo´n MA, Zarini S, Murphy RC (2007) Biosynthesis of eicosanoids and transcellular 5. Fu¨rstenberger G, Krieg P, Mu¨ller-Decker K, Habenicht A (2006) What are cyclooxyge- metabolism of leukotrienes in murine bone marrow cells. J Lipid Res 48:716–725. nases and doing in the driver’s seat of carcinogenesis? Int J Cancer 30. Prakash H, Bala M, Ali A, Goel HC (2005) Modification of gamma radiation induced 119:2247–2254. response of peritoneal macrophages and splenocytes by Hippophae rhamnoides 6. Murphy RC, Gijo´n MA (2007) Biosynthesis and metabolism of leukotrienes. Biochem J (RH-3) in mice. J Pharm Pharmacol 57:1065–1072. 405:379–395. 31. Soule BP, et al. (2007) Effects of gamma radiation on Fc␧RI and TLR-mediated mast cell 7. Steinhilber D (1994) 5-Lipoxygenase: Enzyme expression and regulation of activity. activation. J Immunol 179:3276–3286. Pharm Acta Helv 69:3–14. 32. Whelan J, Golemboski KA, Broughton KS, Kinsella JE, Dietert RR (1997) Characteriza- 8. Rådmark O, Werz O, Steinhilber D, Samuelsson B (2007) 5-Lipoxygenase: Regulation of tion of leukotriene production in vivo and in vitro in resident and elicited peritoneal expression and enzyme activity. Trends Biochem Sci 32:332–341. macrophages in chickens and mice. Leukot Essent Fatty Acids 56:41–49. 9. Clark JD, Schievella AR, Nalefski EA, Lin LL (1995) Cytosolic . J Lipid 33. Kolaczkowska E, Shahzidi S, Seljelid R, van Rooijen N, Plytycz B (2002) Early vascular Mediat Cell Signal 12:83–117. permeability in murine experimental peritonitis is co-mediated by resident peritoneal 10. Peters-Golden M, Brock TG (2000) Intracellular compartmentalization of leukotriene macrophages and mast cells: crucial involvement of macrophage-derived cysteinyl- biosynthesis. Am J Respir Crit Care Med 161:S36–S40. leukotrienes. Inflammation 26:61–71. 11. Woods JW, et al. (1993) 5-lipoxygenase and 5-lipoxygenase-activating protein are 34. Byrum RS, Goulet JL, Snouwaert JN, Griffiths RJ, Koller BH (1999) Determination of the localized in the nuclear envelope of activated human leukocytes. J Exp Med 178:1935– contribution of cysteinyl leukotrienes and leukotriene B4 in acute inflammatory re- 1946. sponses using 5-lipoxygenase- and leukotriene A4 hydrolase-deficient mice. J Immunol 12. Mandal AK, et al. (2008) The nuclear membrane organization of leukotriene synthesis. 163:6810–6819. Proc Natl Acad Sci USA 105:20434–20439. 35. Takeshita K, Sakai K, Bacon KB, Gantner F (2003) Critical role of histamine H4 receptor 13. Yokomizo T, Izumi T, Chang K, Takuwa Y, Shimizu T (1997) A G-protein-coupled in leukotriene B4 production and mast cell-dependent neutrophil recruitment induced receptor for leukotriene B that mediates chemotaxis. Nature 387:620–624. 4 by zymosan in vivo. J Pharmacol Exp Ther 307:1072–1078. 14. Yokomizo T, Kato K, Terawaki K, Izumi T, Shimizu T (2000) A second leukotriene B(4) 36. McCurdy JD, Olynych TJ, Maher LH, Marshall JS (2003) Cutting edge: Distinct Toll-like receptor, BLT2. A new therapeutic target in inflammation and immunological disor- receptor 2 activators selectively induce different classes of mediator production from ders. J Exp Med 192:421–432. human mast cells. J Immunol 170:1625–1629. 15. Lynch KR, et al. (1999) Characterization of the human cysteinyl leukotriene CysLT1 37. Borgeat P, Samuelsson B (1979) Arachidonic acid metabolism in polymorphonuclear receptor. Nature 399:789–793. leukocytes: Unstable intermediate in formation of dihydroxy acids. Proc Natl Acad Sci 16. Heise CE, et al. (2000) Characterization of the human cysteinyl leukotriene 2 receptor. USA 76:3213–3217. J Biol Chem 275:30531–30536. 38. Jakobsson PJ, Mancini JA, Ford-Hutchinson AW (1996) Identification and character- 17. Ciana P, et al. (2006) The orphan receptor GPR17 identified as a new dual uracil nucleotides/cysteinyl-leukotrienes receptor. EMBO J 25:4615–4627. ization of a novel human microsomal glutathione S-transferase with leukotriene C4 18. Maekawa A, Kanaoka Y, Xing W, Austen KF (2008) Functional recognition of a distinct synthase activity and significant sequence identity to 5-lipoxygenase-activating pro- tein and . J Biol Chem 271:22203–22210. receptor preferential for leukotriene E4 in mice lacking the cysteinyl leukotriene 1 and 2 receptors. Proc Natl Acad Sci USA 105:16695–16700. 39. Scoggan KA, Jakobsson PJ, Ford-Hutchinson AW (1997) Production of leukotriene C4 in 19. Funk CD (2005) Leukotriene modifiers as potential therapeutics for cardiovascular different human tissues is attributable to distinct membrane bound biosynthetic disease. Nat Rev Drug Discov 4:664–672. enzymes. J Biol Chem 272:10182–10187. 20. Evans JF, Ferguson AD, Mosley RT, Hutchinson JH (2008) What’s all the FLAP about?: 40. Sala A, Zarini S, Folco G, Murphy RC, Henson PM (1999) Differential metabolism of 5-lipoxygenase-activating protein inhibitors for inflammatory diseases. Trends Phar- exogenous and endogenous arachidonic acid in human neutrophils. J Biol Chem macol Sci 29:72–78. 274:28264–28269. 21. Uhl J, et al. (2002) The 5-lipoxygenase promoter is regulated by DNA methylation. J Biol 41. Di Gennaro A, et al. (2004) Cysteinyl-leukotrienes receptor activation in brain inflam- Chem 277:4374–4379. matory reactions and cerebral edema formation: A role for transcellular biosynthesis 22. Flamand N, Lefebvre J, Surette ME, Picard S, Borgeat P (2006) Arachidonic acid of cysteinyl-leukotrienes. FASEB J 18:842–844. regulates the translocation of 5-lipoxygenase to the nuclear membranes in human 42. Sala A, Bolla M, Zarini S, Mu¨ller-Peddinghaus R, Folco G (1996) Release of leukotriene neutrophils. J Biol Chem 281:129–136. A4 versus leukotriene B4 from human polymorphonuclear leukocytes. J Biol Chem 23. Maclouf JA, Murphy RC (1988) Transcellular metabolism of neutrophil-derived leuko- 271:17944–17948. triene A4 by human platelets. A potential cellular source of leukotriene C4. J Biol Chem 43. Goulet JL, Snouwaert JN, Latour AM, Coffman TM, Koller BH (1994) Altered inflam- 263:174–181. matory responses in leukotriene-deficient mice. Proc Natl Acad Sci USA 91:12852– 24. Folco G, Murphy RC (2006) Eicosanoid transcellular biosynthesis: From cell-cell inter- 12856. actions to in vivo tissue responses. Pharmacol Rev 58:375–388. 44. Kanaoka Y, Maekawa A, Penrose JF, Austen KF, Lam BK (2001) Attenuated zymosan- 25. Fitzpatrick FA, Morton DR, Wynalda MA (1982) Albumin stabilizes leukotriene A4. J Biol induced peritoneal vascular permeability and IgE-dependent passive cutaneous ana- Chem 257:4680–4683. phylaxis in mice lacking leukotriene C4 synthase. J Biol Chem 276:22608–22613.

6of6 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0903851106 Zarini et al. Downloaded by guest on September 27, 2021