Prostaglandin D2 production of prostaglandin H synthase
and identification within canine mast cell granule
PAUL S. THOMAS, AUNDREA N. WILSON, RANDALL E. SCHRECK, AND STEPHEN C. LAZARUS Cardiovascular Research Institute and Department University of California, San Francisco, California Thomas, Paul S., Aundrea N. Wilson, Randall E. Schreck, and Stephen C. Lazarus. Prostaglandin D2 production and identification of prostaglandin H synthase within canine mast cell granule. Am. J. Physiol. 263 (Lung Cell. Mol. Physiol. 7): L168-L176,1992.-We have identified the presence of functional prostaglandin H synthase (PGH synthase, E.C. 1.14.99.1, or cyclooxygenase)within canine mast cell granules by demonstratingthe generationof prostaglandin (PG) D2 from isolated and purified granules incubated with substrate as arachidonic acid or stimulated with calcium ionophore, A23187. This confirms the presenceof both enzyme and substratewithin the granule. Localization of PGH synthase to the granule was confirmed by immunoblotting of the pure granule preparation and by immunocytochemistry usingthe whole cell. In functional studies, colchicine, a microtubule polymerization inhibitor, causeda fall of up to 70%, both in the amount of histamine releasedand in the amount of PGD, generated.This suggests either that functional PGH synthase is closely associatedand coactivated with granulesor that there is an independent association of this enzyme with the microtubule system. Releaseof the preformed and newly formed mediators of the mast cell appear to be closely linked, and prevention of degranulation may therefore attenuate the effects of both classesof mediators. cyclooxygenase; prostaglandins; arachidonic acid; calcium ionophore; colchicine SYNCHRONOUS RELEASE of preformed mediators, such as histamine, and newly formed prostanoid mediators from the mast cell (31) led us to study whether mast cell granules contain the machinery to generate cyclooxygenase products. The products of the cyclooxygenase pathway have many diverse effects, particularly in relation to inflammation. In the lungs and airways, prostaglandin D2 (PGD2) induces bronchoconstriction in dog and human (10, 40) and is generated in response to IgE-mediated activation of mast cells (11, 27, 41). PGDz also has proinflammatory effects, such as neutrophil chemokinesis (12), and has been shown to cause accumulation of eosinophils in the tracheal lumen (8). Cellular membranes are a rich source of arachidonic acid, which is formed from phospholipid by the action of phospholipases. For this reason it was logical to assume that the enzymes of the lipoxygenase and cyclooxygenase pathways would be plasma membrane bound so that they would be in close proximity to substrate and the extracellular release of their products would be facilitated. The use of antibodies directed against enzymes of the cyclooxygenase pathway has not confirmed this view (2% Cyclooxygenase, also known as prostaglandin (PG) H synthase (or PGG/H synthase), catalyzes the conversion of arachidonic acid to a cyclic endoperoxide and adds a further oxygen molecule to form PGGB. PGH THE
L168
1040-0605/92
$2.00
Copyright
of Medicine, 94143
synthase also converts this product to PGH2 by hydroperoxidase activity. PGH synthase has been localized immunocytochemically to the membranes of the endoplasmic reticulum and nucleus in Swiss 3T3 fibroblasts (29), and PGI synthase was found in nuclear and plasma membrane in vascular smooth muscle in several species (34). In this latter study there was PGI synthase immunoreactivity on the plasma membrane of some cells, but, surprisingly, the PGH and PGI synthases were rarely found in the same location, yet the product of PGH synthase, PGH2, serves as substrate for PGI synthase (34). One possible explanation may be that levels of PGH synthase on the plasma membrane are below the limit of detection by immunocytochemistry. Both PGH and PGI synthases are postulated to span membranes with their active sites on the cytoplasmic, COOH-terminal portion of the molecule (13, 22). The activity of the PGH synthase is inhibited by nonsteroidal antiinflammatory drugs which competitively inhibit the binding of substrate to enzyme. In addition, aspirin has been shown to irreversibly acetylate serine at position 530 so that new enzyme must be synthesized before activity is regained (5). PGH2, the product of PGH synthase, is also the substrate for PGD synthetase. This latter enzyme is believed to be cytosolic (35, 39), whereas some workers have reported other enzymes of the cyclooxygenase pathway to be membrane bound (23, 28, 29, 34). How the PGH2 and PGD synthetase interact is unexplained. Two possibilities are enzyme translocation and movement of PGHz itself. Mast cell granules are surrounded by perigranular phospholipid membranes which fuse with the plasma membrane on activation of the cell. There are several ways in which granule exocytosis could be linked to the eicosanoid pathways. Granules contain nonbilayer phospholipid which may be available for the generation of the eicosanoids (4). This led us to the hypothesis that on activation of the cell, these granules could increase the substrate available for phospholipid metabolism. If the enzymes are membrane bound then fusion of granule and plasma membranes would increase the surface area from which the eicosanoid enzymes could generate and release their products into the extracellular space. Eicosanoid production could precede exocytosis so that eicosanoids are secreted into the granule and released along with the preformed mediators. Alternatively, eicosanoid formation could parallel the fusion between granule and plasma membrane but be activated by a separate intracellular messenger system. A further speculation would be that fusion of these two membranes could activate the phospholipases releasing arachidonic
0 1992 the American
Physiological
Society
Downloaded from www.physiology.org/journal/ajplung by ${individualUser.givenNames} ${individualUser.surname} (129.059.095.115) on January 13, 2019.
CYCLOOXYGENASE
ACTIVITY
IN
acid and act as a stimulus for eicosanoid formation. It has been proposed that the active site of PGH synthase is on the cytoplasmic side of the membrane-bound enzyme (13, 22), but the mechanism for the release of the eicosanoids from the cytoplasm of the cell is unknown. Our purpose in this study was to examine the possibility that mast cell granules could generate PGDz along with the preformed mediators. We chose to measure PGDB, since this is the major cyclooxygenase product of the mast cell and has biological relevance in the airway. To examine PGD2 production by mast cell granules, we utilized two preparations currently established in our laboratories. Isolated mast cell granules were studied for their ability to generate PGDz and for the presence of PGH synthase, the first enzyme in the cyclooxygenase cascade leading to synthesis of PGD2. In parallel experiments we used intact mast cells to localize PGH synthase by immunocytochemistry and for pharmacological studies linking PGDz generation to the mast cell granule. These studies are complementary and provide strong evidence for the presence of cyclooxygenase activity within the mast cell granules. METHODS
Cells and subcellular preparations. We have previously reported long-term culture of canine mastocytoma cells (17) and the technique for isolating and purifying intact mast cell granules (38). We studied granulesobtained from two dog mastocytoma cell lines designated BR and G, which have been well characterized in terms of their morphologicaland pharmacological status (17). Cellsof both lines are known to generatePGD, and, in addition, G cellsare able to produce leukotriene C, and 5hydroxy-6,8,11,14-eicosatetraenoic acid (5HETE) (18). The granulepreparations are free from contamination by mitochondria, nuclei, cell cytoplasm, and cell membranes(36-38). Reagents. Unless specified, reagents were obtained from Sigma Chemical (St. Louis, MO) and cell culture media from University of California San Francisco Cell Culture Facility. Animals. The experimental protocols were developed in accordance with the Revised Public Health Service Policy on Humane Care and Use of Laboratory Animals, the published “Guiding Principles in the Care and Use of Animals” approved by the Council of the American Physiological Society and a specific protocol approved by the Committee on Animal Care of the University of California, San Francisco. Mice were obtained and housedby the Animal Care Facility, University of California, San Francisco. Mastocytoma cells. Canine mastocytomasfrom two cell lines, BR and G, have been serially passagedin BALB/c athymic, nude mice. The tumors were developed and disaggregatedas previously described (17) Briefly, tumors were minced into small fragments and digested with collagenaseand deoxyribonuclease(DNase) (Calbiochem,La Jolla, CA). Freshly disaggregated cells or cellskept in short-term suspension(50% Dulbecco’s modified Eagle’s/Hl6 medium containing 25 mM 2v-2hydroxyethylpiperazine-N’-2-ethanesulfonic acid (HEPES), 100,000U/l penicillin, 50 pg/l gentamicin, 250 mg/l histidine, and 2 mM glutamine) were identified by metachromatic staining with toluidine blue and viability assessed by exclusion of erythrosin B vital stain. All preparations were >99% mast cells and those of 93% cell disruption as assessed by erythrosin B vital dye exclusion (3638). A continuous gradient of 0.63-2.63 M sucrosein 4% Ficoll, IO mM HEPES, pH 7.4, wasconstructed usinga gradient mixer and peristaltic pump (Haake-Buchler Autodensiflow IIC, Lenexa, KS). The density increasedfrom 1.04 g/ml at the top to 1.29 g/ml at the bottom of the gradient. The organelle-rich supernatant from the 300 g spin was layered on top of the gradient and centrifuged at 82,000g for 12 h. The granuleband was isolated at a density of 1.24-1.26 g/ml and processedfor further study. In a previous study we have shown that this is uncontaminated by either whole cells or other organelles,as judged by both biochemical and morphologicalcriteria (36-38). Microsomal preparations. Cells were washed in calcium-, magnesium-free(CMF) phosphate-buffered saline (PBS) and resuspendedat a density of 1 x IO7cells/ml in CMF PBS with 2 mM EDTA. The cellswere disrupted with a Branson Sonifier cell disruptor (model 200, Branson Sonic Power Co, Danbury, CT), setting 2, 40% duty cycle and received pulsatile bursts for 60 s. Cell debris was separatedout asa pellet at 10,000g for 15 min and the microsomalfraction collected asa pellet after centrifugation at 100,000g for 60 min. The pellet was resuspended in PBS (containing calcium and magnesium)for studies of PGDz generation. Biochemical studies. Isolated granuleswere washedtwice by centrifugation at 2,000 g in distilled water to remove sucrose. The remaining pellet was then resuspendedin Hanks’ buffered salt solution with 3 g/l bovine gammaglobulin. The granule material was incubated either with calcium ionophore A23187 (Calbiochem,0.3-3 PM) or arachidonic acid (Nu-Check-Prep, Elysian, MN; l-64 PM) at 37°C for 30 min in a final volume of 1 ml. The reaction wasterminated by 5-min incubation at 4”C, followed by centrifugation at 300 g for 10 min. PGD2 released into the supernatant was determined by radioimmunoassay. The effect of cyclooxygenaseinhibitors on the granulepreparation was studied by preincubating granuleswith indomethacin (1 and 10 PM) for 30 min before stimulation with arachidonic acid. BecausePGH, can be degradednonenzymatically to a combination of PGE,, PGD,, and HHT (24), we alsousedreversedphasehigh-pressureliquid chromatography (RP-HPLC) to analyze the granule-derived supernatant. Samplesprepared as above were extracted with Sep-Pak C,, cartridges and analyzed with a Hewlett-Packard liquid chromatograph (model 1090, Hewlett-Packard, Palo Alto, CA). Separation of arachidonic acid metabolites was achieved with a 5 pm-particle octadecylsilaneprecolumn (4.6 mm X 4.5 cm, Altex, Berkeley, CA) and a 5 pm-particle octadecylsilanecolumn (4.6 mm X 25 cm, Microsorb, Rainin Instruments, Emeryville, CA), developedat a flow rate of 1.0 ml/min with a mobile phaseof 30% acetonitrile (Fisher Scientific) and deionizeddistilled water (adjustedto pH 3.5 with glacial acetic acid). By useof this program and monitoring absorbanceat 210 nm with a diode array detector fitted with a flow cell (Hewlett-Packard model 1040A), PGD, and PGE, are separately resolvedat ~26 and 31 min, respectively. To study the amount of enzyme associatedwith granules compared with other organelles,the sucrosegradient for preparing granuleswas fractionated into 22 serial 1.5-ml aliquots, and 0.1 ml of eachaliquot wasincubated in 3.2 PM arachidonic acid as in Biochemical studies. After demonstrating the production of PGD:, by granulesin responseto both arachidonic acid and to A23187, we sought to define whether the enzymes responsiblewere localized to the perigranular membraneor to the
Downloaded from www.physiology.org/journal/ajplung by ${individualUser.givenNames} ${individualUser.surname} (129.059.095.115) on January 13, 2019.
L170
CYCLOOXYGENASE
ACTIVITY
granule matrix. We therefore took isolated purified granule preparations, disrupted them by sonication, and separated the granule matrix from the perigranular membrane fraction by high-speed centrifugation (110,000 g for 1 h). The pellet so derived was designated the perigranular membrane, and both fractions were stimulated with arachidonic acid to determine where the greater activity resided. A further approach to this question was undertaken by stimulating 1 x lo6 whole BR cells with A23187 in a two-stage release experiment (19). Cells were incubated in CMF PBS with 3 PM A23187 at 4°C for 20 min, washed three times in CMF PBS, and incubated at 37°C for 30 min in PBS containing calcium. The preparation was then centrifuged and separated into supernatant containing exocytosed granule matrix material and pellet containing the whole cells. Arachidonic acid (final concentration 19.2 ,uM) was added to each fraction to determine whether there was any PGH synthase and PGD synthase activity in the exocytosed granular material and whether there was any residual activity left in the cells after the stimulation by A23187. To assess the role of the mast cell granule in the generation of the prostanoids, additional experiments were designed to see if antagonizing the polymerization of the microtubular system, so preventing exocytosis, had an effect on the generation of PGD2. Studies of colchicine in human lung mast cells have demonstrated that concentrations of 0.1-l mM are required for inhibition of exocytosis (20). In preliminary concentration-response experiments, we found that 1 mM colchicine partially inhibited histamine release from mastocytoma cells without cytotoxicity. To examine the effect on PGD, production we preincubated cells for 2 h with either 1 mM colchicine or vehicle (0.1% ethanol), after which the cells were washed and additional colchitine or vehicle added at the time of incubation with a concentration range of A23187. In parallel experiments, microsomes were incubated with colchicine or vehicle and arachidonic acid to test whether colchicine had a direct effect on the cyclooxygenase enzyme. Assays. PGD, was determined by a 3H PGD, competitive binding assay using rabbit anti-PGD,, which has negligible cross-reactivity with other cyclooxygenase metabolites (8). Protein assays were by the method of Bradford modified for microassay (Biorad, Richmond, CA) (2). Histamine was determined by the o-phthalaldehyde spectrofluorometric procedure modified for use with an autoanalyzer (33). Immunoblot. Proteins were extracted from sonicated samples of the whole cells or the purified granule fraction by resuspension in 0.1% 3- [ (3-cholamidopropyl)-dimethylammonio] -l-propanesulfonate (CHAPS) with 0.27 mg/ml aprotinin, 0.83 mg/ml chymostatin, and 0.3 mg/ml soybean trypsin inhibitor, solubilized and denatured according to the method of Lammelli and electrophoresed on a 10% sodium dodecylsufate polyacrylamide gel (16). Proteins were transferred to nitrocellulose paper by electroblotting overnight in 30% methanol, 30 mM tris(hydroxymethyl)aminomethane (Tris) base, and 0.22 M glycine at 30 V. The blots were incubated in a blocking solution of 5% bovine serum albumin (BSA) in Tris-buffered saline (TBS) for I h, followed by overnight incubation in 1% BSA TBS with a 1: 100 dilution of polyclonal rabbit antiprostaglandin H synthase antibody (PG20, Oxford Biomedical Research, Oxford, MI). The blots were then incubated in a l:3,000 dilution of goat anti-rabbit biotinylated conjugate (Bio-Rad) in TBS followed by a further incubation in l:l,OOO avidin-alkaline phosphatase conjugate (Extravidin). Between each incubation the blots were washed once in 1% Tween 20 in TBS then washed twice in TBS. The blots were finally developed by incubation in 1 ml of 0.1% nitroblue tetrazolium in 0.05% sodium borate pH 9.6,9 ml of Verona1 acetate buffer, 20 ml of 2 M magnesium chloride, and 100 ml of 1 mg/ml indolyl phosphate. Quench controls were run as above but incubated with antibody (PG20) preincubated with
IN PURE
MAST
CELL
GRANULES
7 mg of purified PGH synthase (PGOI, Oxford Biomedical Research) per milliliter of antibody. Other controls consisted of omitting either primary or secondary antibody and the substitution of nonimmune normal rabbit serum for the primary antibody. Immunocytochemistry. Approximately 1 x IO5 freshly disaggregated G cells were cytocentrifuged onto slides, fixed in icecold acetone, incubated with Chromotrope 2R, 0.03% at 37°C for 10 min, and washed in PBS. The cells were then incubated with either mouse monoclonal anti-prostaglandin H synthase (PG2l: cyo-1 and PG22: cyo-5, Oxford Biomedical) or rabbit polyclonal anti-prostaglandin H synthase antibody (PG20) diluted 1:1 with PBS containing 2% normal goat serum (NGS, Cappel, Malvern, PA) and 0.6% Triton X-100 for 4 h. The slides were then washed three times with PBS plus 0.3% Triton X-100,1% NGS. Goat anti-mouse IgG-FITC (Cappel) at a 1:40 dilution in PBS plus 0.3% Triton X-100, 1% NGS was then added for 20 min. After a final two washes in PBS the cells were mounted in glycerol. The slides were viewed using a Zeiss fluorescence microscope with epifluorescence optics and photographs taken on Ektachrome 400 (Kodak, Rochester, NY). Controls consisted of omitting the primary antibody, omitting the secondary antibody, use of normal rabbit serum (preimmune), and preincubation of the primary antibody for 20 min with purified PGH synthase as above. Electron immunocytochemistry. Because the PGH synthase is exquisitely sensitive to fixation and embedding procedures (29)) we utilized high-pressure quick-freezing followed by freeze substitution to prepare cells for electron immunocytochemistry (9). Mastocytoma cells were suspended with high-molecular-weight dextran in culture media and quick frozen in a Balzer HPM 010 high-pressure freezer (Bal-tee, Middlebury, CT). Cells were freeze-substituted at -90°C for 3 days in a Reichert CS Auto freeze dryer (Cambridge Instruments, Cambridge, MA) with acetone and 1% osmium tetroxide. Cells were brought slowly to -20°C and osmium tetroxide removed by three washes with acetone. Cells were infiltrated with LR White resin (Polysciences, Warrington, PA) at room temperature and polymerized at 50°C for 24 h. Thin sections were cut with a diamond knife and placed on nickel grids. The grids were then incubated on a 20 ~1 drop of PBS containing 0.8% BSA and 0.1% cold-water fish gelatin for IO min to block nonspecific binding. This solution was used throughout as a wash and diluent. The grid was then transferred to 20 ~1 of rabbit anti-PGH synthase antibody (PG20, diluted 1:l) and incubated overnight at 4°C. It was washed five times and placed on a 20-~1 drop of mouse antirabbit IgG (Accurate Chemical & Scientific, Westbury, NY, diluted 1:50) for 1 h. The grid was then washed five times and placed on a 20-~1 drop of 5- nm gold-conjugated goat anti-mouse IgG diluted 1:lO (EY Laboratories, San Mateo, CA). Grids were washed five times in double-distilled water, after which they were fixed in 2% glutaraldehyde, incubated with 2% osmium tetroxide and 2% uranyl acetate, coated with a Formvar film (E. F. Fulham, Schenectady, NY), and viewed on a Zeiss EM IO electron microscope. Controls included omission of primary antibody (rabbit anti-PGH synthase), omission of mouse antirabbit IgG, and normal rabbit serum. Statistics. Statistical analysis was by analysis of variance (ANOVA) and Student’s t test where appropriate (Abacus Concepts, Berkeley, CA), and results are expressed as means t SE unless indicated otherwise. RESULTS
PGDz generation from isolated and purified granules.
There was a significant concentration-related rise in PGD2 generation from granules isolated and purified from both BR and G cell lines in response to exogenous
Downloaded from www.physiology.org/journal/ajplung by ${individualUser.givenNames} ${individualUser.surname} (129.059.095.115) on January 13, 2019.
CYCLOOXYGENASE
ACTIVITY
IN PURE MAST
CELL GRANULES
Ll71
arachidonic acid (Fig. 1). Calcium ionophore, A23187, also caused a small but significant rise in the production of PGDz (Fig. 2). When activated granules were analyzed 3 by reverse-phase high-performance liquid chromatograPGD, phy (RP-HPLC), using a solvent system that resolves (PSYKI 0 -I PGD2 and PGEB we found a single peak coeluting with protein) L NT authentic PGDz and no evidence for PGE2. This suggests k that the PGD2 measured in the granule preparations was 1 1 "+A generated enzymatically and was not a nonenzymatic de0 .Ol .I 1 IO composition product of PGHz (24). A231 87 (PM) Whole cells generated PGD2 in response to both 2. PGD, generation by isolated and purified mast cell granules from A23187 and to arachidonic acid. In response to 1 PM Fig. BR cell line. Granules were incubated for 30 min with a concentration A23187, BR cells produced 7.7 t 2.3 ng PGD&g protein range of calcium ionophore A23187 (mean s;t SE, n. = 3, P < 0.05, (mean t SD, n = 3) and G cells made 2.0 t 0.1 ng ANOVA). PGD,/pg protein; in response to 32 PM arachidonic 1001 acid, BR cells generated 10.4 t 2.9 ng PGD,/pg protein and G cells generated 9.3 & 2.4 ng PGD,/pg protein. The G cells generate a larger amount of PGD2 per cell but on a per unit protein basis, which normalizes for the differPGD2 50 ence in cell size between each line, the BR cell line ap(PmI peared to have slightly greater ability to generate this protein) prostaglandin. Serial fractions down the sucrose density gradient were assayed for the ability of various other organelles to generate PGD2. A large amount of activity was localized to 0 20 40 60 80 fractions conta ining a mixture of cytosol and broken Arachidonic acid (PM) membranes. Of the fractions on the gradient containing Fig. 3. Effect of indomethacin on generation of PGDz by isolated and the subcellular organelles, the granule fraction contained purified BR mast cell granules. Granules were preincubated with indomethacin (1 PM, closed squares, 10 PM closed circles) or with buffer 25.0 t 2.9% of the total activity (n = 2). alone (open circles), then incubated with a concentration range of Indomethacin inhibited granule-associated cyclooxygenase. Preincubation of granules with indomethacin, 10 arachidonic acid (mean t SE, n = 3, P < 0.05, ANOVA). PM, resulted in a 65.3 t 5.6% fall in the amount of PGDz produced in response to 64 PM arachidonic acid (Fig. 3, arachidonic acid. There was no discernable increase in P < 0.05, ANOVA) PGDz formation after the addition of arachidonic acid to When purified granules were separated by sonication the degranulation supernatant, implying that the active and high-speed centrifugation into membrane and nonenzymes required for the formation of PGD2 were not membrane fractions, only the nonmembrane fraction produced PGD, (1.2 ng PGDJml) in response to 32 PM present in the exocytosed material (control: 0.97 t 0.11; arachidonic acid, 19.2 PM: 0.98 t 0.13 ng PGD,/lg proarachidonic acid. Because these studies and previous imtein, n = 9). Incubation of the cell pellet with the-same munocytochemistry su.ggested that cyclooxygen ase activity may be localized to the mast cell granule, we the refore dose of arachidonic acid resulted in the generation of 3.9 sought to determine whether active enzyme was released + 0.54 ng PGD,/pg protein vs. a control value of 0.81 t from whole cells on degranulation. After degranulation by 0.12. This latter experiment demonstrates that the twoa two-stage activation with A23187, degranulation super- stage activation with A23187 does not inactivate cell bound cyclooxygenase activity. natant and the remaining cell pellet were incubated with Effect of colchicine. Those cells preincubated with 1 mM colchicine showed a 37.5-59.6% reduction in ionophore-induced histamine release and a 57.3-75.6% reduction in the formation of PGDz (Fig. 4, P < 0.05, ANOVA). Cell viability analysis by erythrocin B vital dye and lactate dehydrogenase release revealed no difference between the control cells and those incubated with colchitine, nor did colchicine cause degranulation. There was
and identification within canine mast cell granule
PAUL S. THOMAS, AUNDREA N. WILSON, RANDALL E. SCHRECK, AND STEPHEN C. LAZARUS Cardiovascular Research Institute and Department University of California, San Francisco, California Thomas, Paul S., Aundrea N. Wilson, Randall E. Schreck, and Stephen C. Lazarus. Prostaglandin D2 production and identification of prostaglandin H synthase within canine mast cell granule. Am. J. Physiol. 263 (Lung Cell. Mol. Physiol. 7): L168-L176,1992.-We have identified the presence of functional prostaglandin H synthase (PGH synthase, E.C. 1.14.99.1, or cyclooxygenase)within canine mast cell granules by demonstratingthe generationof prostaglandin (PG) D2 from isolated and purified granules incubated with substrate as arachidonic acid or stimulated with calcium ionophore, A23187. This confirms the presenceof both enzyme and substratewithin the granule. Localization of PGH synthase to the granule was confirmed by immunoblotting of the pure granule preparation and by immunocytochemistry usingthe whole cell. In functional studies, colchicine, a microtubule polymerization inhibitor, causeda fall of up to 70%, both in the amount of histamine releasedand in the amount of PGD, generated.This suggests either that functional PGH synthase is closely associatedand coactivated with granulesor that there is an independent association of this enzyme with the microtubule system. Releaseof the preformed and newly formed mediators of the mast cell appear to be closely linked, and prevention of degranulation may therefore attenuate the effects of both classesof mediators. cyclooxygenase; prostaglandins; arachidonic acid; calcium ionophore; colchicine SYNCHRONOUS RELEASE of preformed mediators, such as histamine, and newly formed prostanoid mediators from the mast cell (31) led us to study whether mast cell granules contain the machinery to generate cyclooxygenase products. The products of the cyclooxygenase pathway have many diverse effects, particularly in relation to inflammation. In the lungs and airways, prostaglandin D2 (PGD2) induces bronchoconstriction in dog and human (10, 40) and is generated in response to IgE-mediated activation of mast cells (11, 27, 41). PGDz also has proinflammatory effects, such as neutrophil chemokinesis (12), and has been shown to cause accumulation of eosinophils in the tracheal lumen (8). Cellular membranes are a rich source of arachidonic acid, which is formed from phospholipid by the action of phospholipases. For this reason it was logical to assume that the enzymes of the lipoxygenase and cyclooxygenase pathways would be plasma membrane bound so that they would be in close proximity to substrate and the extracellular release of their products would be facilitated. The use of antibodies directed against enzymes of the cyclooxygenase pathway has not confirmed this view (2% Cyclooxygenase, also known as prostaglandin (PG) H synthase (or PGG/H synthase), catalyzes the conversion of arachidonic acid to a cyclic endoperoxide and adds a further oxygen molecule to form PGGB. PGH THE
L168
1040-0605/92
$2.00
Copyright
of Medicine, 94143
synthase also converts this product to PGH2 by hydroperoxidase activity. PGH synthase has been localized immunocytochemically to the membranes of the endoplasmic reticulum and nucleus in Swiss 3T3 fibroblasts (29), and PGI synthase was found in nuclear and plasma membrane in vascular smooth muscle in several species (34). In this latter study there was PGI synthase immunoreactivity on the plasma membrane of some cells, but, surprisingly, the PGH and PGI synthases were rarely found in the same location, yet the product of PGH synthase, PGH2, serves as substrate for PGI synthase (34). One possible explanation may be that levels of PGH synthase on the plasma membrane are below the limit of detection by immunocytochemistry. Both PGH and PGI synthases are postulated to span membranes with their active sites on the cytoplasmic, COOH-terminal portion of the molecule (13, 22). The activity of the PGH synthase is inhibited by nonsteroidal antiinflammatory drugs which competitively inhibit the binding of substrate to enzyme. In addition, aspirin has been shown to irreversibly acetylate serine at position 530 so that new enzyme must be synthesized before activity is regained (5). PGH2, the product of PGH synthase, is also the substrate for PGD synthetase. This latter enzyme is believed to be cytosolic (35, 39), whereas some workers have reported other enzymes of the cyclooxygenase pathway to be membrane bound (23, 28, 29, 34). How the PGH2 and PGD synthetase interact is unexplained. Two possibilities are enzyme translocation and movement of PGHz itself. Mast cell granules are surrounded by perigranular phospholipid membranes which fuse with the plasma membrane on activation of the cell. There are several ways in which granule exocytosis could be linked to the eicosanoid pathways. Granules contain nonbilayer phospholipid which may be available for the generation of the eicosanoids (4). This led us to the hypothesis that on activation of the cell, these granules could increase the substrate available for phospholipid metabolism. If the enzymes are membrane bound then fusion of granule and plasma membranes would increase the surface area from which the eicosanoid enzymes could generate and release their products into the extracellular space. Eicosanoid production could precede exocytosis so that eicosanoids are secreted into the granule and released along with the preformed mediators. Alternatively, eicosanoid formation could parallel the fusion between granule and plasma membrane but be activated by a separate intracellular messenger system. A further speculation would be that fusion of these two membranes could activate the phospholipases releasing arachidonic
0 1992 the American
Physiological
Society
Downloaded from www.physiology.org/journal/ajplung by ${individualUser.givenNames} ${individualUser.surname} (129.059.095.115) on January 13, 2019.
CYCLOOXYGENASE
ACTIVITY
IN
acid and act as a stimulus for eicosanoid formation. It has been proposed that the active site of PGH synthase is on the cytoplasmic side of the membrane-bound enzyme (13, 22), but the mechanism for the release of the eicosanoids from the cytoplasm of the cell is unknown. Our purpose in this study was to examine the possibility that mast cell granules could generate PGDz along with the preformed mediators. We chose to measure PGDB, since this is the major cyclooxygenase product of the mast cell and has biological relevance in the airway. To examine PGD2 production by mast cell granules, we utilized two preparations currently established in our laboratories. Isolated mast cell granules were studied for their ability to generate PGDz and for the presence of PGH synthase, the first enzyme in the cyclooxygenase cascade leading to synthesis of PGD2. In parallel experiments we used intact mast cells to localize PGH synthase by immunocytochemistry and for pharmacological studies linking PGDz generation to the mast cell granule. These studies are complementary and provide strong evidence for the presence of cyclooxygenase activity within the mast cell granules. METHODS
Cells and subcellular preparations. We have previously reported long-term culture of canine mastocytoma cells (17) and the technique for isolating and purifying intact mast cell granules (38). We studied granulesobtained from two dog mastocytoma cell lines designated BR and G, which have been well characterized in terms of their morphologicaland pharmacological status (17). Cellsof both lines are known to generatePGD, and, in addition, G cellsare able to produce leukotriene C, and 5hydroxy-6,8,11,14-eicosatetraenoic acid (5HETE) (18). The granulepreparations are free from contamination by mitochondria, nuclei, cell cytoplasm, and cell membranes(36-38). Reagents. Unless specified, reagents were obtained from Sigma Chemical (St. Louis, MO) and cell culture media from University of California San Francisco Cell Culture Facility. Animals. The experimental protocols were developed in accordance with the Revised Public Health Service Policy on Humane Care and Use of Laboratory Animals, the published “Guiding Principles in the Care and Use of Animals” approved by the Council of the American Physiological Society and a specific protocol approved by the Committee on Animal Care of the University of California, San Francisco. Mice were obtained and housedby the Animal Care Facility, University of California, San Francisco. Mastocytoma cells. Canine mastocytomasfrom two cell lines, BR and G, have been serially passagedin BALB/c athymic, nude mice. The tumors were developed and disaggregatedas previously described (17) Briefly, tumors were minced into small fragments and digested with collagenaseand deoxyribonuclease(DNase) (Calbiochem,La Jolla, CA). Freshly disaggregated cells or cellskept in short-term suspension(50% Dulbecco’s modified Eagle’s/Hl6 medium containing 25 mM 2v-2hydroxyethylpiperazine-N’-2-ethanesulfonic acid (HEPES), 100,000U/l penicillin, 50 pg/l gentamicin, 250 mg/l histidine, and 2 mM glutamine) were identified by metachromatic staining with toluidine blue and viability assessed by exclusion of erythrosin B vital stain. All preparations were >99% mast cells and those of 93% cell disruption as assessed by erythrosin B vital dye exclusion (3638). A continuous gradient of 0.63-2.63 M sucrosein 4% Ficoll, IO mM HEPES, pH 7.4, wasconstructed usinga gradient mixer and peristaltic pump (Haake-Buchler Autodensiflow IIC, Lenexa, KS). The density increasedfrom 1.04 g/ml at the top to 1.29 g/ml at the bottom of the gradient. The organelle-rich supernatant from the 300 g spin was layered on top of the gradient and centrifuged at 82,000g for 12 h. The granuleband was isolated at a density of 1.24-1.26 g/ml and processedfor further study. In a previous study we have shown that this is uncontaminated by either whole cells or other organelles,as judged by both biochemical and morphologicalcriteria (36-38). Microsomal preparations. Cells were washed in calcium-, magnesium-free(CMF) phosphate-buffered saline (PBS) and resuspendedat a density of 1 x IO7cells/ml in CMF PBS with 2 mM EDTA. The cellswere disrupted with a Branson Sonifier cell disruptor (model 200, Branson Sonic Power Co, Danbury, CT), setting 2, 40% duty cycle and received pulsatile bursts for 60 s. Cell debris was separatedout asa pellet at 10,000g for 15 min and the microsomalfraction collected asa pellet after centrifugation at 100,000g for 60 min. The pellet was resuspended in PBS (containing calcium and magnesium)for studies of PGDz generation. Biochemical studies. Isolated granuleswere washedtwice by centrifugation at 2,000 g in distilled water to remove sucrose. The remaining pellet was then resuspendedin Hanks’ buffered salt solution with 3 g/l bovine gammaglobulin. The granule material was incubated either with calcium ionophore A23187 (Calbiochem,0.3-3 PM) or arachidonic acid (Nu-Check-Prep, Elysian, MN; l-64 PM) at 37°C for 30 min in a final volume of 1 ml. The reaction wasterminated by 5-min incubation at 4”C, followed by centrifugation at 300 g for 10 min. PGD2 released into the supernatant was determined by radioimmunoassay. The effect of cyclooxygenaseinhibitors on the granulepreparation was studied by preincubating granuleswith indomethacin (1 and 10 PM) for 30 min before stimulation with arachidonic acid. BecausePGH, can be degradednonenzymatically to a combination of PGE,, PGD,, and HHT (24), we alsousedreversedphasehigh-pressureliquid chromatography (RP-HPLC) to analyze the granule-derived supernatant. Samplesprepared as above were extracted with Sep-Pak C,, cartridges and analyzed with a Hewlett-Packard liquid chromatograph (model 1090, Hewlett-Packard, Palo Alto, CA). Separation of arachidonic acid metabolites was achieved with a 5 pm-particle octadecylsilaneprecolumn (4.6 mm X 4.5 cm, Altex, Berkeley, CA) and a 5 pm-particle octadecylsilanecolumn (4.6 mm X 25 cm, Microsorb, Rainin Instruments, Emeryville, CA), developedat a flow rate of 1.0 ml/min with a mobile phaseof 30% acetonitrile (Fisher Scientific) and deionizeddistilled water (adjustedto pH 3.5 with glacial acetic acid). By useof this program and monitoring absorbanceat 210 nm with a diode array detector fitted with a flow cell (Hewlett-Packard model 1040A), PGD, and PGE, are separately resolvedat ~26 and 31 min, respectively. To study the amount of enzyme associatedwith granules compared with other organelles,the sucrosegradient for preparing granuleswas fractionated into 22 serial 1.5-ml aliquots, and 0.1 ml of eachaliquot wasincubated in 3.2 PM arachidonic acid as in Biochemical studies. After demonstrating the production of PGD:, by granulesin responseto both arachidonic acid and to A23187, we sought to define whether the enzymes responsiblewere localized to the perigranular membraneor to the
Downloaded from www.physiology.org/journal/ajplung by ${individualUser.givenNames} ${individualUser.surname} (129.059.095.115) on January 13, 2019.
L170
CYCLOOXYGENASE
ACTIVITY
granule matrix. We therefore took isolated purified granule preparations, disrupted them by sonication, and separated the granule matrix from the perigranular membrane fraction by high-speed centrifugation (110,000 g for 1 h). The pellet so derived was designated the perigranular membrane, and both fractions were stimulated with arachidonic acid to determine where the greater activity resided. A further approach to this question was undertaken by stimulating 1 x lo6 whole BR cells with A23187 in a two-stage release experiment (19). Cells were incubated in CMF PBS with 3 PM A23187 at 4°C for 20 min, washed three times in CMF PBS, and incubated at 37°C for 30 min in PBS containing calcium. The preparation was then centrifuged and separated into supernatant containing exocytosed granule matrix material and pellet containing the whole cells. Arachidonic acid (final concentration 19.2 ,uM) was added to each fraction to determine whether there was any PGH synthase and PGD synthase activity in the exocytosed granular material and whether there was any residual activity left in the cells after the stimulation by A23187. To assess the role of the mast cell granule in the generation of the prostanoids, additional experiments were designed to see if antagonizing the polymerization of the microtubular system, so preventing exocytosis, had an effect on the generation of PGD2. Studies of colchicine in human lung mast cells have demonstrated that concentrations of 0.1-l mM are required for inhibition of exocytosis (20). In preliminary concentration-response experiments, we found that 1 mM colchicine partially inhibited histamine release from mastocytoma cells without cytotoxicity. To examine the effect on PGD, production we preincubated cells for 2 h with either 1 mM colchicine or vehicle (0.1% ethanol), after which the cells were washed and additional colchitine or vehicle added at the time of incubation with a concentration range of A23187. In parallel experiments, microsomes were incubated with colchicine or vehicle and arachidonic acid to test whether colchicine had a direct effect on the cyclooxygenase enzyme. Assays. PGD, was determined by a 3H PGD, competitive binding assay using rabbit anti-PGD,, which has negligible cross-reactivity with other cyclooxygenase metabolites (8). Protein assays were by the method of Bradford modified for microassay (Biorad, Richmond, CA) (2). Histamine was determined by the o-phthalaldehyde spectrofluorometric procedure modified for use with an autoanalyzer (33). Immunoblot. Proteins were extracted from sonicated samples of the whole cells or the purified granule fraction by resuspension in 0.1% 3- [ (3-cholamidopropyl)-dimethylammonio] -l-propanesulfonate (CHAPS) with 0.27 mg/ml aprotinin, 0.83 mg/ml chymostatin, and 0.3 mg/ml soybean trypsin inhibitor, solubilized and denatured according to the method of Lammelli and electrophoresed on a 10% sodium dodecylsufate polyacrylamide gel (16). Proteins were transferred to nitrocellulose paper by electroblotting overnight in 30% methanol, 30 mM tris(hydroxymethyl)aminomethane (Tris) base, and 0.22 M glycine at 30 V. The blots were incubated in a blocking solution of 5% bovine serum albumin (BSA) in Tris-buffered saline (TBS) for I h, followed by overnight incubation in 1% BSA TBS with a 1: 100 dilution of polyclonal rabbit antiprostaglandin H synthase antibody (PG20, Oxford Biomedical Research, Oxford, MI). The blots were then incubated in a l:3,000 dilution of goat anti-rabbit biotinylated conjugate (Bio-Rad) in TBS followed by a further incubation in l:l,OOO avidin-alkaline phosphatase conjugate (Extravidin). Between each incubation the blots were washed once in 1% Tween 20 in TBS then washed twice in TBS. The blots were finally developed by incubation in 1 ml of 0.1% nitroblue tetrazolium in 0.05% sodium borate pH 9.6,9 ml of Verona1 acetate buffer, 20 ml of 2 M magnesium chloride, and 100 ml of 1 mg/ml indolyl phosphate. Quench controls were run as above but incubated with antibody (PG20) preincubated with
IN PURE
MAST
CELL
GRANULES
7 mg of purified PGH synthase (PGOI, Oxford Biomedical Research) per milliliter of antibody. Other controls consisted of omitting either primary or secondary antibody and the substitution of nonimmune normal rabbit serum for the primary antibody. Immunocytochemistry. Approximately 1 x IO5 freshly disaggregated G cells were cytocentrifuged onto slides, fixed in icecold acetone, incubated with Chromotrope 2R, 0.03% at 37°C for 10 min, and washed in PBS. The cells were then incubated with either mouse monoclonal anti-prostaglandin H synthase (PG2l: cyo-1 and PG22: cyo-5, Oxford Biomedical) or rabbit polyclonal anti-prostaglandin H synthase antibody (PG20) diluted 1:1 with PBS containing 2% normal goat serum (NGS, Cappel, Malvern, PA) and 0.6% Triton X-100 for 4 h. The slides were then washed three times with PBS plus 0.3% Triton X-100,1% NGS. Goat anti-mouse IgG-FITC (Cappel) at a 1:40 dilution in PBS plus 0.3% Triton X-100, 1% NGS was then added for 20 min. After a final two washes in PBS the cells were mounted in glycerol. The slides were viewed using a Zeiss fluorescence microscope with epifluorescence optics and photographs taken on Ektachrome 400 (Kodak, Rochester, NY). Controls consisted of omitting the primary antibody, omitting the secondary antibody, use of normal rabbit serum (preimmune), and preincubation of the primary antibody for 20 min with purified PGH synthase as above. Electron immunocytochemistry. Because the PGH synthase is exquisitely sensitive to fixation and embedding procedures (29)) we utilized high-pressure quick-freezing followed by freeze substitution to prepare cells for electron immunocytochemistry (9). Mastocytoma cells were suspended with high-molecular-weight dextran in culture media and quick frozen in a Balzer HPM 010 high-pressure freezer (Bal-tee, Middlebury, CT). Cells were freeze-substituted at -90°C for 3 days in a Reichert CS Auto freeze dryer (Cambridge Instruments, Cambridge, MA) with acetone and 1% osmium tetroxide. Cells were brought slowly to -20°C and osmium tetroxide removed by three washes with acetone. Cells were infiltrated with LR White resin (Polysciences, Warrington, PA) at room temperature and polymerized at 50°C for 24 h. Thin sections were cut with a diamond knife and placed on nickel grids. The grids were then incubated on a 20 ~1 drop of PBS containing 0.8% BSA and 0.1% cold-water fish gelatin for IO min to block nonspecific binding. This solution was used throughout as a wash and diluent. The grid was then transferred to 20 ~1 of rabbit anti-PGH synthase antibody (PG20, diluted 1:l) and incubated overnight at 4°C. It was washed five times and placed on a 20-~1 drop of mouse antirabbit IgG (Accurate Chemical & Scientific, Westbury, NY, diluted 1:50) for 1 h. The grid was then washed five times and placed on a 20-~1 drop of 5- nm gold-conjugated goat anti-mouse IgG diluted 1:lO (EY Laboratories, San Mateo, CA). Grids were washed five times in double-distilled water, after which they were fixed in 2% glutaraldehyde, incubated with 2% osmium tetroxide and 2% uranyl acetate, coated with a Formvar film (E. F. Fulham, Schenectady, NY), and viewed on a Zeiss EM IO electron microscope. Controls included omission of primary antibody (rabbit anti-PGH synthase), omission of mouse antirabbit IgG, and normal rabbit serum. Statistics. Statistical analysis was by analysis of variance (ANOVA) and Student’s t test where appropriate (Abacus Concepts, Berkeley, CA), and results are expressed as means t SE unless indicated otherwise. RESULTS
PGDz generation from isolated and purified granules.
There was a significant concentration-related rise in PGD2 generation from granules isolated and purified from both BR and G cell lines in response to exogenous
Downloaded from www.physiology.org/journal/ajplung by ${individualUser.givenNames} ${individualUser.surname} (129.059.095.115) on January 13, 2019.
CYCLOOXYGENASE
ACTIVITY
IN PURE MAST
CELL GRANULES
Ll71
arachidonic acid (Fig. 1). Calcium ionophore, A23187, also caused a small but significant rise in the production of PGDz (Fig. 2). When activated granules were analyzed 3 by reverse-phase high-performance liquid chromatograPGD, phy (RP-HPLC), using a solvent system that resolves (PSYKI 0 -I PGD2 and PGEB we found a single peak coeluting with protein) L NT authentic PGDz and no evidence for PGE2. This suggests k that the PGD2 measured in the granule preparations was 1 1 "+A generated enzymatically and was not a nonenzymatic de0 .Ol .I 1 IO composition product of PGHz (24). A231 87 (PM) Whole cells generated PGD2 in response to both 2. PGD, generation by isolated and purified mast cell granules from A23187 and to arachidonic acid. In response to 1 PM Fig. BR cell line. Granules were incubated for 30 min with a concentration A23187, BR cells produced 7.7 t 2.3 ng PGD&g protein range of calcium ionophore A23187 (mean s;t SE, n. = 3, P < 0.05, (mean t SD, n = 3) and G cells made 2.0 t 0.1 ng ANOVA). PGD,/pg protein; in response to 32 PM arachidonic 1001 acid, BR cells generated 10.4 t 2.9 ng PGD,/pg protein and G cells generated 9.3 & 2.4 ng PGD,/pg protein. The G cells generate a larger amount of PGD2 per cell but on a per unit protein basis, which normalizes for the differPGD2 50 ence in cell size between each line, the BR cell line ap(PmI peared to have slightly greater ability to generate this protein) prostaglandin. Serial fractions down the sucrose density gradient were assayed for the ability of various other organelles to generate PGD2. A large amount of activity was localized to 0 20 40 60 80 fractions conta ining a mixture of cytosol and broken Arachidonic acid (PM) membranes. Of the fractions on the gradient containing Fig. 3. Effect of indomethacin on generation of PGDz by isolated and the subcellular organelles, the granule fraction contained purified BR mast cell granules. Granules were preincubated with indomethacin (1 PM, closed squares, 10 PM closed circles) or with buffer 25.0 t 2.9% of the total activity (n = 2). alone (open circles), then incubated with a concentration range of Indomethacin inhibited granule-associated cyclooxygenase. Preincubation of granules with indomethacin, 10 arachidonic acid (mean t SE, n = 3, P < 0.05, ANOVA). PM, resulted in a 65.3 t 5.6% fall in the amount of PGDz produced in response to 64 PM arachidonic acid (Fig. 3, arachidonic acid. There was no discernable increase in P < 0.05, ANOVA) PGDz formation after the addition of arachidonic acid to When purified granules were separated by sonication the degranulation supernatant, implying that the active and high-speed centrifugation into membrane and nonenzymes required for the formation of PGD2 were not membrane fractions, only the nonmembrane fraction produced PGD, (1.2 ng PGDJml) in response to 32 PM present in the exocytosed material (control: 0.97 t 0.11; arachidonic acid, 19.2 PM: 0.98 t 0.13 ng PGD,/lg proarachidonic acid. Because these studies and previous imtein, n = 9). Incubation of the cell pellet with the-same munocytochemistry su.ggested that cyclooxygen ase activity may be localized to the mast cell granule, we the refore dose of arachidonic acid resulted in the generation of 3.9 sought to determine whether active enzyme was released + 0.54 ng PGD,/pg protein vs. a control value of 0.81 t from whole cells on degranulation. After degranulation by 0.12. This latter experiment demonstrates that the twoa two-stage activation with A23187, degranulation super- stage activation with A23187 does not inactivate cell bound cyclooxygenase activity. natant and the remaining cell pellet were incubated with Effect of colchicine. Those cells preincubated with 1 mM colchicine showed a 37.5-59.6% reduction in ionophore-induced histamine release and a 57.3-75.6% reduction in the formation of PGDz (Fig. 4, P < 0.05, ANOVA). Cell viability analysis by erythrocin B vital dye and lactate dehydrogenase release revealed no difference between the control cells and those incubated with colchitine, nor did colchicine cause degranulation. There was
