395

Biochimica et Biophysicu Acta, 1042 (1990) 395-403 Elsevier

BBALIP

53323

Altered arachidonic acid metabolism during differentiation of the human monoblastoid cell line U937 L. Koehler ‘, R. Hass ‘, K. Wessel ‘, D.L. Dewitt and M. Goppelt-Struebe

(Received

Key words:

Cell differentiation;

Arachidonic

19 September

2, V. Kaever ‘, K. Resch

’

’

1989)

acid metabolism; Icosanoid; (Human U937 cell line)

Cyclooxygenase;

Phospholipase

A*;

The human cell line U937 was used as a model for differentiation along the mononuclear phagocyte lineage. Following treatment with the phorbol ester TPA, PGE, and TxB, secretion was induced ~-lo-fold, and both PGF,, and PGI, levels became detectable in the supematant of TPA-differentiated U937 cells. The content of the prostaglandin precursor, arachidonic acid, remained unchanged in the cellular phospholipids of undifferentiated and TPA-differentiated U937 cells. Of the enzymes involved in the availability and metabolism of arachidonic acid, phospholipase A, activity was increased tfold in the membranes of TPA-differentiated U937 cells, whereas lysophosphatide acyltransferase activity remained unaltered. Cyclooxygenase activity, however, was enhanced 5-lO-fold, which was due to enhanced expression of the enzyme as demonstrated by dot-blot analysis. The data suggest that the capacity to secrete prostaglandins is acquired during differentiation with TPA and results mainly from an increased cyclooxygenase activity. Despite the capacity of TPA-differentiated U937 cells to synthesize prostaglandins, none of the known monocytic stimuli further stimulated prostaglandin secretion in TPA-differentiated U937 cells. Generation of leukotrienes appears to represent a later state in the differentia~on along the mon~yte-macrophage lineage, since neither LTB, nor cysteinyl-leukotrienes were detectable in the supematants of either undifferentiated or TPA-differentiated U937 cells.

Introduction It is well established that activated human monocytes are capable of metabolizing arachidonic acid via both cyclooxygenase and lipoxygenase pathways [l]. The intracellular level of arachidonic acid is regulated by activities of the liberating phospholipases and the reacylating enzyme, lysophosphatide acyltransferase [2]. Recently published data suggest that the ability to synthesize prostaglandins from liberated arachidonic acid might reflect a function of cellular differentiation. The enhanced capacity to secrete prostaglandins was

Abbreviations: PGE,, prostagIandin E,; TxB,, thromboxan B,; PGIZ, prostacyclin; PGF,,, prostaglandin Fzn; LTB.,, leukotriene B4; LTC,/D,/E,, leukotriene C,/D,/E,; TPA, 12-O-tetradecanoylphorbol 13-acetate. Correspondence: M. Goppelt-Stntebe, macology, Medical School Hannover, 3000 Hannover 61, F.R.G. 00052760/90/$03.50

Division of Molecular Konstanty-Gutschow-Strasse

0 1990 Elsevier Science Publishers

Phar8,

B.V. (Biomedical

related to an induction of cyclooxygenase, the initial enzyme metabolizing arachidonic acid. Thus, epidermal growth factor induced cyclooxygenase in human amnion cells [3] and mouse osteoblastic cells [4], platelet-derived growth factor in Swiss 3T3 cells [5] and transforming growth factor-p in mouse osteoblastic cells [6]. In addition, cyclooxygenase was induced by hormones in rat ovarian follicles [7]. As a model for investigating monocytic differentiation, tumor cell lines are widely used which acquire monocytic characteristics following treatment with various agents. The human histiocytic cell line U937 first established by Sundstriim and Nilsson [8] was characterized as a progenitor cell line in the monocytic lineage, U937 cells can be induced to differentiate along the monocyte-macrophage lineage by incubation with various agents including dimethyl sulfoxide, retinoic acid, 12-0-tetradecanoylphorbol13-acetate (TPA) and several metabolites of vitamin D [9,10]. TPA-induced differentiation results in cessation of proliferation and significant alterations in the morphology of the cells. U937 Division)

396 cells, which when growing in suspension have a smooth and round surface, become adherent by forming cell clusters and extend pseudopodia after phorbol ester treatment [11,12]. The morphological changes observed during the differentiation process suggest functional changes of the U937 cells. With respect to arachidonic acid metabolism, undifferentiated U937 cells constitutively secrete very low amounts of prostaglandins. In contrast to monocytes, prostaglandin synthesis cannot be increased by any of the common stimuli such as calcium ionophore and zymosan. However, exogenously-added arachidonic acid enhanced PGE, secretion [13]. TPA-differentiated U937 cells have been shown to be more active in terms of PGE, synthesis after incubation with lipopolysaccharide from Escherichia coli [14]. Leukotriene production has not yet been unequivocally defined in either cell type. Compared to other models of differentiation, little is known about the enzymes involved in arachidonic acid metabolism in the U937 cells. Our results show that the 50-100-fold induction of prostanoid synthesis during differentiation with TPA is a result of increased activities of both phospholipase A, and cyclooxygenase. However, enhancement of cyclooxygenase activity was more pronounced suggesting a central role of this enzyme in the capacity of these cells to metabolize arachidonic acid with resulting prostanoid synthesis.

Materials and Methods

Culture of U937 cell line U937 cells were obtained from the American Type Culture Collection (Rockville, MD, U.S.A.). The cells were cultured and differentiated as described elsewhere

illI. Determination of cellular non-esterified fatty acids and fatty acid bound to phospholipids Cellular lipids were extracted as described previously [15]. Phospholipids and non-esterified fatty acids were separated from other lipids by thin-layer chromatography on silica glass plates (Schleicher and Schuell, Basel, F.R.G.) using a solvent system of petroleum ether/diethyl ether/glacial acetic acid (50 : 50 : 1, v/v). Phospholipids and non-esterified fatty acids were identified by comparison with authentic standards and scraped off. Phospholipids were transmethylated with sodium methylate in the presence of silica gel. By this method only acyl-linked fatty acids were hydrolized and methylated [16]. Non-esterified fatty acids were methylated directly by the reaction with diazomethane generated from N-methyl-N-nitroso-p-toluene sulfonamide (Aldrich, Steinheim, F.R.G.). Fatty acid methyl ester

were analyzed

by capillary

gas-liquid

chromatography

[171.

Release of eicosanoids U937 cells and TPA-differentiated cells were incubated overnight with RPM1 1640 medium, containing 1% heat-inactivated human-AB-serum in a flat bottom microtiter plate (96 wells) or 24-well plates (Nunc, Wiesbaden, F.R.G.) at a cell density of 5. lo5 cells/ml. Prostaglandin generation. Prostaglandin synthesis was induced by incubation with arachidonic acid (5 . 10 ‘-2 . 10e5 M) (NuCheck Prep., Elysian, MN, U.S.A.) or by stimulation of the cells with calcium ionophore A23187 (l-50 pg/ml), unopsonized zymosan (16-1600 pg/ml), TPA (lo-‘, 10-s M) (all from Sigma, Deisenhofen, F.R.G.), lipopolysaccharide from E. coli 0127 : B7 (0.1-50 pg/ml) (Difco Laboratories, Detroit, MI, U.S.A.), y-interferon (102-lo5 U/ml), recombinant human interleukin-lb (20-40 U/ml), recombinant murine interleukin-la (l-10 U/ml) (both from Biogen Res. Geneva, Switzerland) or recombinant human tumor necrosis factor-a (1 to lo5 U/ml) (BASF, Ludwigshafen, F.R.G.). The stimuli were diluted in RPM1 1640 medium containing 1% human AB serum. Aliquots of the supernatants were taken at the time points indicated. Concentrations of prostaglandins including PGE,, TxB,, the immediate nonenzymatic conversion product of thromboxane A,, PGF2a and 6-keto-PGF,,, the hydrolysis product of prostacyclin, were determined by radioimmunoassay (RIA) using a modified doubleantibody method as described previously [18]. The specific antibodies were raised in rabbits. The crossreactivities of the antibodies directed against PGE, or TxBz to other prostaglandins were less than 0.01% [18]. The crossreactivity of the PGF,, antibody against PGE,, TxB,, PGD, or 6-keto-PGF,,, was less than O.l%, O.l%, 2.5% and l%, respectively (unpublished results). None of the antibodies recognized arachidonic acid at the concentrations used. Leukotriene generation. For leukotriene generation the cells were challenged with either ionophore A23187 (0.1-50 pg/ml), arachidonic acid (5 . lop6 M), a combination of ionophore and arachidonic acid, unopsonized zymosan (160 pgg/ml) or lipopolysaccharide from E. coli (10 pg/ml) for 1 and 24 h. For determination of the cysteinyl-leukotrienes, two different antibodies were used. The polyclonal rabbit antibody was specific for LTC, since the crossreactivity to LTD, and LTE, was less than 4% and O.l%, respectively [18]. The monoclonal antibody recognized LTC, to the same extent as its degradation products LTD, and LTE, (the antibody was kindly provided by Dr. M. Reinke, Pharmakologisches Institut der Universitat Erlangen, F.R.G.) [19]. The LTB,-RIA (NEN, Boston, MA, U.S.A.) and the monoclonal cysteinyl-leukotriene RIA were performed

397 using a charcoal separation limit was 150 pg/ml.

technique.

The

detection

High performance liquid chromatography (HPLC) TPA-differentiated U937 cells (1 * lo6 cells) were incubated overnight with 1 PCi [‘4C]arachidonic acid (specific activity 56 mCi/mmol) (NEN, Boston, MA, U.S.A.). After washing with medium containing 0.5% bovine serum albumin, the cells were incubated with unlabeled arachidonic acid (2. 10e5 M) for 24 h to enhance arachidonic acid metabolism. Supernatants were collected and eicosanoids extracted with octadecylsilyl-Bond Elute columns (Analytichem, Harbor City, CA, U.S.A.) for HPLC analysis (LC 41, Brucker Franzen, Bremen, F.R.G.). The column (MZ-Analysentechnik, Mainz, F.R.G.) was 250 mm length and 1.6 mm i.d., filled with Nucleosil 5 pm ODS particles. For separation of the whole eicosanoid spectrum, a threestep gradient with two different mobile phases was applied. Phase (A) consisted of water/ acetonitrile/ acetic acid (68.0 : 31.9 : 0.1, v/v) adjusted to pH 4.5 with sodium acetate. Phase (B) was composed of methanol/ acetonitrile/ acetic acid (60.0 : 39.9 : 0.1, v/v). The gradient was performed as followed: O-5 min phase (A), 5-10 min a linear gradient to a final composition of 56% (A)/44% (B), lo-20 min isocratic conditions, 20-40 min a linear gradient from 56% (A)/44% (B), to finally 100% (B). These conditions were held constant up to the end of the chromatogram at 75 min. The flow rate was set to 0.2 ml/mm. Rialuma scintillation liquid (Baker, Gross-Gerau, F.R.G.) was mixed to the eluent with 1 ml/mm. Radioactivity was monitored with an HPLC-P-radiation counter (Ramona, D., Raytest, Straubenhardt, F.R.G.). The quantities of arachidonic acid metabolites released were too small to be detected by optical methods. CeN preparation for enzyme assays Adherent TPA-differentiated cells were washed twice with phosphate-buffered saline (PBS) and harvested with a rubber policeman in 100 mM N-tris(hydroxymethyl)methylglycine buffer (Tricine) (Serva, Heidelberg, F.R.G.) (pH 8.5) from culture dishes. In suspension growing undifferentiated U937 cells were centrifuged at 600 X g for 10 min and washed twice in PBS. The pellets of both undifferentiated and TPA-differentiated U937 cells were resuspended in Tricine buffer and sonicated three times for 10 s at 50W (Labsonic 1510, Braun, Melsungen, F.R.G.). After centrifugation for 10 min at 600 x g the supernatant was taken off and centrifuged 30 min at 100000 X g. The resulting pellet containing the cell membranes was resuspended in Tritine buffer. The protein concentration of this crude membrane fraction was determined by a micro Bradford assay [20] using bovine serum albumin as standard protein. The crude membrane faction was used for

phospholipase cyclooxygenase

A z, lysophosphatide enzyme assays.

acyltransferase

and

Determination of phospholipase A, activity Phospholipase A, was determined as originally described by Flesch and Ferber [21] with minor modifications. The assay is based on the liberation of 14Clabeled arachidonic acid from a-1-palmitoyl-2arachidonoylphosphatidylcholine (NEN, Boston, MA, U.S.A.) in the presence of 1 mM CaCl,. 4 mM EDTA were added to control tubes characterizing the calcium dependency of the phospholipase A,. The reaction was started by addition of 1 PM labeled phospholipid (25 000 cpm) in a total volume of 250 ~1. The mixture was incubated at 37” C in glass tubes with shaking for 30 min. The reaction was stopped by adding 375 ~1 of ice-cold 2-propanol/l M HCI (1 : 0.086, v/v). Liberated [ l4 Clarachidonic acid was separated from phospholipids by heptane extraction and measured in a liquid scintillation counter. Specific activity of phospholipase A, (pmol/mg protein per min) was calculated as follows: Spec. act. =

2(cpm(sample,

-cpmchlank)

cpmC,tandardj

)

pmol substrate min. (mg protein)

wherecPm(s,nlple)

is the measured radioactivity of the hydrolysis of the subsamp1e; cpm(hlank), spontaneous strate; cpm(,,,,,,,,,, total radioactivity; and pmol substrate, [‘4C]arachidonylphosphatidylcholine in pmol. Determination of lysophosphatide acyltransferase activity Lysophosphatide acyltransferase activity was measured as previously described [22]. Briefly, the acylation of l-acyl-glycerophosphatidylcholine with arachidonoylCoA was observed by the reaction of liberated Coenzyme A with the dye 5,5-dithio-bis(nitrobenzoic acid) (Sigma, Deisenhofen, F.R.G.). The absorbance was determined in a Gilford spectrophotometer (Corning Lab. Sciences Company) at 412 nm at 37’C. Lysophosphatide acyltransferase activity was calculated from the initial velocity of the reaction (molar absorption coefficient 13600 M-’ cm-‘). Determination of cyclooxygenase/ PGE, isomerase activity The enzyme assay was performed similarly as described by Kosaka et al. [23]. Protein of the 100000 X g pellet was resuspended and diluted at concentrations indicated in 50 mM Hepes/l40 mM KC1 buffer (pH 8.0). The assay mixture containing 3.75 mM of reduced glutathione (Boehringer, Mannheim, F.R.G.), 8 PM hemoglobin, 320 PM tryptophan (Sigma, Deisenhofen, F.R.G.) and protein of the cell membranes in Hepes/KCl buffer was preincubated for 10 min at 37 o C. Diclofenac (Sigma, Deisenhofen, F.R.G.), an inhibitor of the cyclooxygenase, was included in control

398 tubes at 250 PM. The enzyme assay was started by addition of arachidonic acid (lo-’ to 1O-4 M). The total volume consisted of 200 ~1. The reaction was stopped by freezing the assay mixture immediately in liquid nitrogen. Before measuring the content of PGE, by radioimmunoassay the tubes were centrifuged at 16 000 x g for 1 min. No crossreactivity of the PGE, antibody was measured to any constituents of the assay. Cyclooxygenase dot-blot analysis Undifferentiated and TPA-differentiated U937 cells were harvested and the crude membrane fraction prepared as described above. The protein was diluted in PBS and applied to Immobilon-P transfer membrane (Millipore, Bedford, U.S.A.) using a dot-blot apparatus (Bio-Rad, Richmond, U.S.A.). The polyclonal rabbit anti-cyclooxygenase antibody was diluted as described previously [7] in the blocking buffer which consisted of PBS, 10 mg/ml human IgG (Sandoz, Nlirnberg, F.R.G.) and 5% low-fat dry milk (w/v) (pH 7.4). The detection limit of the antibody was 25 ng of cyclooxygenase [7]. Antibodies bound to cyclooxygenase were detected by biotinylated goat anti-rabbit IgG and visualized by streptavidin-peroxidase complex (both from Amersham, Braunschweig, F.R.G.) using the method of Ref. [24]. Results

Prostaglandin secretion Prostaglandin release was measured in the cell culture supernatant of undifferentiated and TPA-differentiated U937 cells. Undifferentiated U937 cells produced barely detectable amounts of PGE,, but could be induced to secrete PGE, by incubation with arachidonic acid in a concentration-dependent manner (Fig. la). In contrast, TPA-differentiated U937 cells spontaneously secreted marked levels of PGE,. The prostaglandin secretion achieved by exogenous arachidonic acid was enhanced 50-100-fold after differentiation with TPA

TABLE

I

Prostagiandin cells

secretion of undifferentmted

and TPA-differentiated

U937

Undifferentiated and TPA-differentiated cells were incubated with 2.10m5 M arachidonic acid for 24 h. Supernatants were collected and the content of prostaglandins determined by radioimmunoassay as described in Methods. Data are the meansfS.D. of three separate experiments each performed in quadruplicates and given in pg/ml per 5.105 cells per 24 h. Prostaglandin

Undifferentiated U937 cells

PGE, unstimulated + arachidonic TxB, unstimulated + arachidonic

TPA-Differentiated U937 cells

cells acid

15* 13 468 li_379

21161 12040*

cells acid

75* 43 479*117

6006* 2362 4026Ok13660

PGF,, unstimulated + arachidonic

cells acid

< 20 77i

PGI, unstimulated + arachidonic

cells acid

< 20 < 20

459i 305li

3

48+ 424+

806 1906

244 740 4 317

(Fig. lb). The secretion of other prostaglandins including TxB,, PGI, and PGF,, is shown in Table I. Undifferentiated U937 cells were capable of metabolizing arachidonic acid to PGE, and TxB,. Two further prostaglandins, PGF,, and PGI,, were only detectable in the supernatant of TPA-differentiated U937 cells. The results obtained by radioimmunoassay were confirmed by high-performance liquid chromatography analysis (HPLC). The HPLC system used separated prostaglandins, leukotrienes, hydroxyeicosatetraenoic acids (HETEs) and arachidonic acid. The secreted prostaglandins eluted at retention times consistent with added standards (Fig. 2). The major arachidonic acid metabolites in TPA-differentiated U937 cells were PGE, and TxB,, followed by PGF,, and PGI,. No evidence

la

01, 01

4

8

24 T,me

Fig. 1. (a) Time course of PGE, secretion of undifferentiated M (0) and 2.10m5 M (H) arachidonic acid, After the times Data are the means+ SD. of quadruplicate determinations. ated U937 cells were incubated without (A) or with 5.10-” supernatants were collected and PGE, determined by

(h)

U937 cells, Undifferentiated U937 cells were incubated without (A) or with 1 .10m5 indicated, supernatants were collected and PGE, determined by radioimmunoassay. (b) Time course of PGE, secretion of TPA-differentiated U937 cells. TPA-differentiM (0). 1.10-s M (0) and 2.10m5 M (W) arachidonic acid. After the times indicated. radioimmunoassay. Data are the means f SD. of quadruplicate determinations.

399 LT

PG I

t

HeteS

ARA

I-II-

_/

40 60 (mln) Fig. 2. HPLC analysis of the supernatant of TPA-differentiated lJ937 cells. TPA-differentiated U937 cells were incubated overnight with 1 PCi [‘4C]arachidonic acid. After washing with medium containing 0.5% bovine serum albumin the cells were incubated with 2.10-’ M arachidonic acid for 24 h. Eicosanoids (PC, prostaglandins; LT. leukotrienes; HETEs, hydroxyeicosatetraenoic acids; ABA, arachidonit acid) were extracted and analyzed by HPLC as described in Methods. The peaks were identified by the retention times of authentic standards and were 6-keto-PGF,, (I), TxB, (2). PGF,, (3). PGE, (4) and arachidonic acid (5).

0

20

Time

for the synthesis of leukotrienes or HETEs was obtained. Various stimuli were tested for their capacity to enhance prostaglandin synthesis in TPA-differentiated cells. Depending on the stimulus the incubation time varied from 1 to 48 h for Ca*+-ionophore A23187, TPA, zymosan and lipopolysaccharide from E. coli and from 1 to 5 days for interleukin-1 and tumor necrosis factor. None of these agents under varying conditions showed any stimulation of prostaglandin synthesis in TPA-differentiated U937 cells.

In a separate experiment, cyclooxygenase activity was blocked by 10u5 M acetyl salycilic acid, which completely inhibited the prostaglandin synthesis. Again, no leukotrienes were detected under those experimental conditions (data not shown). Composition of cellular non-esterified fatty acids and fatty acids bound to cellular phospholipids Lipids were extracted from undifferentiated and TPA-differentiated U937 cells. Phospholipids and nonesterified fatty acids were separated from other lipids by thin-layer chromatography and the composition of non-esterified fatty acids and fatty acids of the phospholipids determined by capillary gas-liquid chromatography. Only traces of non-esterified palmitic acid and stearic acid were observed in undifferentiated and TPA-differentiated U937 cells. The detection limit was about 60 nM fatty acid/60 PM phospholipid (data not shown). The ratio of phospholipid to protein was not altered during differentiation, thus allowing a direct comparison of fatty acids found in the phospholipids from undifferentiated and TPA-differentiated U937 cells. The content of unsaturated fatty acids was raised in TPA-differentiated U937 cells as indicated by a decreased ratio of saturated to unsaturated fatty acids. However, no significant increase of the precursor of eicosanoid synthesis, arachidonic acid, was measured (Table III). Phospholipase AI actioity The calcium-dependent phospholipase A 2 activity in the membrane fraction was measured in both undifferentiated and TPA-differentiated U937 cells (Table IV). In comparison to undifferentiated U937 cells, membrane-associated phospholipase A 2 activity in TPA-differentiated cells was increased 2-fold.

TABLE

II

Determination of eicosanoidsin the supernatanr of TPA-differentiated Leukotriene synthesis Supernatants of undifferentiated and TPA-differentiated U937 cells were also analyzed for their content of LTB, and cysteinyl-leukotrienes. To detect possible degradation products of LTC,, two antibodies were used in the radioimmunoassay (RIA), one with specificity only for LTC,, the other one for LTC,, LTD, and LTE,. The detection limit for both was 150 pg/ml in RIA measurements. The cells were stimulated to secrete leukotrienes by ionophore A23187, free arachidonic acid, a combination of ionophore and arachidonic acid, unopsonized zymosan and lipopolysaccharide from E. coli for 1 and 24 h. None of these stimuli could enhance the generation of detectable amounts of leukotrienes (Table II).

lJ93 7 cells TPA-differentiated U937 cells were incubated with medium, arachidonic acid (5.10m6 M), ionophore A23187 (1 pg/ml) and lipopolysaccharide from E. coli (10 pg/ml) for 24 h. Supernatants were removed and eicosanoids determined by radioimmunoassay. Data are the means+S.D. of duplicates with biological relevance given in pg/ml. Detection limit for leukotriene B4 and leukotriene C, was 150 pg/ml. One typical experiment out of three similar ones is shown. n.d.. not detectable. Prostaglandin Control Arachidonic acid Ionophore A23187 Lipopolysaccharide

Leukotriene

Leukotriene

‘52

84

C4

1186k293 9239+ 86 1585 5 239 1437+ 13

n.d. nd. nd. nd.

n.d. n.d. n.d. n.d.

400 TABLE

III

Fatty acid composttion of phospholipids TPA-differentrated U937 cells

from

undifferentiated

and

Lipids were extracted from undifferentiated and TPA-differentiated U937 cells. Phospholipids were separated by thin-layer chromatography. After transmethylation, fatty acid methylesters were analysed by gas chromatography. Data are the means+S.D. of three preparations and given as percent fatty acid. Fatty acid

Undifferentiated U937 cells

TPA-differentiated U937 cells

14:o 16:0 16: 1 18:O 18: 1 1x:2 20:4 22 : 4 22 : 5 22 : 6

2.4kl.O 18.4kO.4 8.8+1.3 15.5 + 1.3 36.4+0.1 1.950.1 7.7 * 0.9 5.750.2 0.8 * 0.1 4.2 * 0.8

0.8kO.3 12.3kl.l 3.7kl.l 16.6k1.5 38.2 + 2.5 2.2 + 0.4 8.4*1.0 8.0 * 0.6 2.6+0.1 5.5 + 0.1

1.9+0.1

1.2kO.l

Ratio Saturated

: unsaturated

Lysophosphatide acyltransferase activity Lysophosphatide acyltransferase activity, catalyzing the reacylation of activated fatty acids into lysophospholipids, was detectable in both undifferentiated and TPA-differentiated U937 cells (Table IV). There was no significant change in enzyme activity during differentiation with TPA. Cyclooxygenase / PGE, isomerase activity The enzyme activity was determined in the membrane fraction by measuring the generation of PGE, from exogenously added arachidonic acid. The conversion was inhibited by more than 90% by the cyclooxygenase inhibitor diclofenac. The protein dependency was linear in the range of 25 pg to 250 pg. A dose dependency revealed the optimal concentratin of 1 10e5

TABLE

Fig. 3. Arachidonic acid dose dependency of cyclooxygenase activity of undifferentiated and TPA-differentiated U937 cells. Crude membranes of undifferentiated (closed bars) and TPA-differentiated U937 cells (open bars) were prepared as described in Methods. Equal amounts of protein were incubated with arachidonic acid at concentrations indicated and the content of PGE, formed determined by radioimmunoassay. Data are means + SD. of quadruplicates and given in ng PGE,/mg protein per min.

M arachidonic acid in both undifferentiated and TPA-differentiated U937 cells (Fig. 3). Measuring a time dependency of cyclooxygenase, PGE, synthesis increased linearly in both undifferentiated and TPA-differentiated U937 cells reaching a plateau after 15 min (Fig. 4). Following 15 min incubation with 1 . lo- 5 M arachidonic acid the cyclooxygenase activity of TPA-differen-

/

-

IV

Enzyme activities in membranes of undifferentiated ated U937 cells involved in prostanoid generation

and TPA-differenti-

Membranes were prepared and enzymes measured as described in Methods. Enzyme activities are given in nmol/mg protein per min for lysophosphatide acyltransferase, in pmol/mg protein per min for phospholipase A z and in ng PGE, /mg protein per min for cyclooxygenase. Data represent the means k SD. of three preparations.

0;

01

5

15

30

Time (mbn)

Lysophosphatide acyltransferase Phospholipase A, Cyclooxygenase

Undifferentiated U937 cells

TPA-Differentiated U937 cells

10.8_+1.1 5.5 i 0.4 0.4 * 0.05

11.0~1.8 10.2 ;t 0.7 2.7 + 0.2

Fig 4. Time course of cyclooxygenase activity of undifferentiated and TPA-differentiated U937 cells. Crude membranes of undifferentiated (0) and TPA-differentiated U937 cells (m) were prepared as described in Methods. Equal amounts of protein were incubated with 1. 10m5 M arachidonic acid for the times indicated and the content of PGE, formed determined by radioimmunoassay. Data are the means * S.D. of quadruplicates and given in ng PGE,/mg protein.

401

25.C

3.1;

1.5E

Fig. 5. Dot-blot analysis of cyclooxygenase in undifferentiated and TPA-differentiated U937 cells. Crude membranes of undifferentiated (A) and TPA-differentiated (B) U937 cells were prepared as described in Methods. Protein was applied to Immobilon blot membrane. After incubation with a polyclonal rabbit anti-cyclooxygenase antibody, immunopositive complexes were visualized by incubation with biotinylated goat anti-rabbit IgG and streptavidin peroxidase complex.

tiated cells was 8.0 &-2.0-fold (n = 6) higher to undifferentiated cells (Table IV).

compared

Dot-blot analysis of cyclooxygenase Crude membranes prepared from undifferentiated and TPA-differentiated U937 cells were examined by dot-blot analysis to determine the expression of cyclooxygenase at the protein level. By comparison of equal amounts of protein from undifferentiated and TPA-differentiated U937 cells, a markedly enhanced expression of cyclooxygenase in TPA-differentiated U937 cells became obvious (Fig. 5). Discussion It is generally accepted that eicosanoids are potent mediators of inflammatory processes [25]. Hence the regulation of eicosanoid synthesis is of great interest. In the last decade, investigations have been focused on the availability of the common precursor free arachidonic acid which depends on the enzymatic release from storage phospholipids and reuptake in this pool by

phospholipase and lysophosphatide acyltransferase, respectively. Recently published data have shown that the cyclooxygenase enzyme system also might play a crucial role for the production of prostaglandins. Goerig et al. have shown that cyclooxygenase is induced during cellular maturation [26] and in addition, growth factors and hormones have been demonstrated to stimulate prostaglandin secretion by de novo synthesis of cyclooxygenase [3-71. In this study, the human monoblastoid cell line U937 was used as a model for macrophage-like differentiation. Undifferentiated U937 cells produced barely detectable amounts of prostaglandins constitutively, but could be stimulated to secrete PGE, and TxB, after incubation with exogenous arachidonic acid. After differentiation of these cells with TPA however, a 50-100fold enhanced prostaglandin secretion either constitutively or in response to arachidonic acid stimulation was observed. In contrast to undifferentiated U937 cells, two further prostaglandins, PGF,, and PGI,, became detectable following TPA-induced differentiation. The content of the prostaglandin precursor arachidonit acid in these cells could not account for the observed induction of prostaglandin synthesis during differentiation since both the level of free arachidonic acid was below the detection limit of about 60 nM fatty acid per 60 PM phospholipid and the amount of arachidonic acid bound to phospholipids was not altered in TPAdifferentiated U937 cells compared to undifferentiated U937 cells. Therefore, alterations of enzymes involved in the prostaglandin pathway had to be considered. The level of arachidonic acid available for eicosanoid synthesis is primarily controlled by the activities of either phospholipase A, cleaving fatty acids from phospholipids and the lysophosphatide acyltransferase, the reacylating enzyme. Besides enhanced liberation of arachidonic acid by phospholipase A 2, reduced lysophosphatide acyltransferase activity has been shown to be an effective mechanism of increasing the pool of free fatty acid and concomitantly eicosanoids [27,31]. In the U937 cell line membrane-associated phospholipase A, activity was increased 2-fold following TPA-induced differentiation, while the activity of the lysophosphatide acyltransferase remained unchanged. These results are in good accordance with the findings of Wiederholt et al. [12], who have shown an approximately 2-fold increase in the release of non-radioactive arachidonic acid after TPAinduced differentiation, documented by gas-liquid chromatography. The changes in membrane-associated phospholipase A, activity, however, cannot fully explain the observed increases in prostaglandin secretion during differentiation since spontaneous PGE, secretion in both undifferentiated and TPA-differentiated U937 cells was induced by factors in excess of 50-fold. In this respect,

402

it is interesting to note that cyclooxygenase/PGE, isomerase activity was enhanced 5-lo-fold in TPA-differentiated U937 cells suggesting that the cyclooxygenase plays a crucial role in the induction of prostaglandin synthesis. This was confirmed by a significantly enhanced expression of cyclooxygenase in TPA-differentiated U937 cells compared to undifferentiated U937 cells as determined by dot-blot analysis. Thus, the finding of both increased membrane-associated phospholipase A 2 and cyclooxygenase activity may explain the enhanced prostaglandin secretion following differentiation with TPA. In HL-60 cells, which have several characteristics in common with U937 cells, an enhanced expression of PGH synthetase was observed during differentiation with TPA [26]. These data suggest that the regulation of the cycloxygenase activity is primarily responsible for the ability to generate prostaglandins at a certain state of maturation. In macrophages, prostaglandin synthesis can be stimulated by various agents [l]. None of these including TPA, ionophore A23187, unopsonized zymosan, concanavalin A (131 and lipopolysaccharide from S. typhimurium [28], have been shown to activate prostaglandin secretion in undifferentiated U937 cells. As the capacity of the prostaglandin synthesis is much higher in TPA-differentiated U937 cells, prostaglandin secretion was attempted to be induced in these cells using ionophore, TPA, zymosan, lipopolysaccharide (from E. cob), y-interferon, Interleukin-1 and tumor necrosis factor. However, even TPA-differentiated cells did not produce significantly higher amounts of prostaglandins from endogenous arachidonic acid compared to unstimulated cells. Similar observations concerning TPA, zymosan and ionophore were obtained previously [13]. These results indicate a major functional defect in the U937 cell line. Alternatively, it is possible that differentiated U937 cells are already stimulated to their maximal extent following treatment with TPA. There is one report by Schenkein [14] that lipopolysaccharide from E. co/i activated PGE, and TxB, synthesis in both cell types. Our experiments could not confirm these results in terms of increased prostaglandin secretion. The reason for this discrepancy is not known. Beside prostaglandins major products of the arachidonic acid metabolism in macrophages are leukotrienes [l]. The release of leukotrienes by U937 cells was suggested by Cobb et al. [13] and Wiederholt et al. [12]. In supernatants of cells, which were prelabeled with arachidonic acid, Cobb et al. [13] and Wiederholt et al. [12] found radioactive peaks, tentatively identified as leukotrienes; the nature of the radioactive peaks however, was not further characterized. In contrast, our experiments showed by two independent methods no capacity of the U937 cell line to synthesize LTB, and cysteinyl-leukotrienes. Using both HPLC and RIA, no leukotrienes were detectable in either unstimulated cells

or in cells stimulated by arachidonic acid, ionophore A23187. zymosan and lipopolysaccharide, although these stimuli have been demonstrated to induce the generation of leukotrienes in human monocytes [29]. Cyclooxygenase inhibitors such as acetyl salycilic acid can increase leukotriene levels in certain cellular systems [30], but even in the presence of this cyclooxygenase inhibitor, no leukotriene synthesis was detectable. Recently we demonstrated by fluorescence activated cell-sorter analysis, that during TPA-induced differentiation, U937 cells do not express certain surface markers associated with mature monocytes/macrophages [ll], suggesting that the lack of the lipoxygenase enzyme system might be due to the immaturity of TPA-differentiated U937 cells. Summarising the data, the capacity to produce significant amounts of prostaglandins is acquired during TPA-induced differentiation of the human monoblastoid cell line U937. The induction of prostaglandin secretion is due to enhanced cyclooxygenase activity concomitant with a small increase in membrane-associated phospholipase A, activity, suggesting that cyclooxygenase plays an important role in the regulation of prostaglandin synthesis. Induced cyclooxygenase activity was due to de novo synthesis of this enzyme during differentiation. In accordance with the surface marker data demonstrating that phorbol ester treatment did not achieve terminally differentiated U937 cells, TPA-induced differentiation did not result in leukotriene secretion, suggesting that the expression of the lipoxygenase enzyme system may represent a late state during monocytic differentiation. Acknowledgements This work was supported by a grant of the Deutsche Forschungsgemeinschaft SFB 244/B5. We thank Mrs. A. Ehlers for her technical assistance. We are indebted to Dr. M. Schulze for his advice in the dot-blot technique. The authors thank Dr. N. Topley for the critical review of the manuscript. References 1 Adams, D.O. and Hamilton, 2 3 4

5

6

T.A. (1984) Annu. Rev. Immunol. 2. 283-318. Irvine, R.F. (1982) Biochem. J. 204, 3-16. Casey, M.L.. Korte, K. and MacDonald, P.C. (1988) J. Biol. Chem. 263, 7846-7854. Yokota, K., Kusaka, M., Ohshima, T., Yamamoto, S., Kurihara. N., Yoshino, T. and Kumegawa. M. (1986) J. Biol. Chem. 261. 15410-15415. Habenicht, A.J.R., Goerig, M., Grulich, J., Rothe, D., Gronwald, R., Loth. U., Schettler, G., Kommerell, B. and Ross, R. (1985) J. Clin. Invest. 75, 1381-387. Sumitani. K.. Kawata. T.. Yoshimoto, T.. Yamamoto. S. and Kumegawa, M. (1989) Arch. Biochem. Biophys. 270, 5X8-595.

403 7 Hedin, L., Gaddy-Kurten, D., Kurten, R., Dewitt, D.L., Smith, W.L. and Richards, J.S. (1987) Endocrinology 121, 722-731. 8 Sundstrom, C. and Nilsson, K. (1976) Int. J. Cancer 17, 565-577. 9 Gildlend, M., Om, A., Pattengale, P.K., Janson, M., Wigzell, H. and Nilsson, K. (1981) Nature 292, 848-850. 10 Dodd, R.C., Cohen, MS., Newman, S.C. and Gray, T.K. (1983) Proc. Natl. Acad. Sci. USA 80, 7538-7541. 11 Hass, R., Barthels, H., Hopley, N., Hadam, M., Koehler, L., Goppelt-Struebe, M. and Resch, K. (1989) Eur. J. Cell Biol. 48, 282-293. 12 Wiederholt, M.D., Anderson, K.M. and Harris, J.E. (1988) Biochim. Biophys. Acta 959, 296-304. 13 Cobb, M.A., Hsueh, W., Pachman, L.M. and Barnes, W.T. (1983) J. Reticuloendothel. Sot. 33, 197-206. 14 Schenkein, H.A. (1986) Cell. Immunol. 102, 307-314. 15 Goppelt, M. and Resch, K. (1984) Anal. Biochem. 140, 152-156. 16 Knapp, D.R. (1979) Handbook of Analytical Derivatization Reaction, p. 164, John Wiley, New York. 17 Goppelt, M. and Resch, K. (1987) B&him. Biophys. Acta 904, 22-28. 18 Kaever, V., Goppelt-Struebe, M. and Resch, K. (1988) Prostaglandins 35, 885-902. 19 Reinke, M. (1989) Naunyn-Schmiedeberg’s Arch. Pharmacol. 339, Suppl., 157.

20 Redinbaugh, M.G. and Campell, W.H. (1985) Anal. Biochem. 147, 144-147. 21 Flesch, J. and Ferber, E. (1986) Biochim. Biophys. Acta 889, 6-14. 22 Goppelt-Struebe, M., Pfannkuche, H.J., Gemsa, D. and Resch, K. (1987) Biochem. J. 247, 773-777. 23 Kosaka, Y., Kawabe, H. and Ishii, A. (1987) Agents Actions 21, 33-37. 24 Peters, H.J., Baumgarten, H. and Schulze, M. (1985) Monoklonale Antikorper: Herstellung und Charakterisierung, Springer, Berlin. 25 Ferreira, S.H. (1981) Trends Pharmacol. Sci. 183-188. 26 Goerig, M., Habenicht, A.J.R., Heitz, R., Zeh, W., Katus, H., Kommerell, B., Ziegler, R. and Glomset, J.A. (1987) J. Clin. Invest. 79, 903-911. 27 Kroner, E.E., Peskar, B.A., Fischer, H. and Ferber, E. (1981) J. Biol. Chem. 256. 3690-3697. 28 Kurland, J.I., Pelus, L.M., Ralph, P., Bockman, R.S. and Moore, M.A.S. (1979) Proc. Natl. Acad. Sci. USA 76, 2326-2330. 29 Bigby. T.D. and Holtzman, M.J. (1987) J. Immunol. 138, 1546-1550. 30 Brune, K., Aehringhaus, U. and Peskar, B. (1984) Agents Actions 14, 729-734. 31 Goppelt-Struebe, M., Koermer, C.-F., Hausmann, G.. Gemsa, D. and Resch, K. (1986) Prostaglandins 32, 373-385.