- Email: [email protected]
Prostacyclin biosynthesis in vascular endothelium is not inhibited by cyclic AMP. Studies with 3-isobutyl-1-methylxanthine and forskolin
THROMBOSIS RESEARCH 28; 637-647, 1982 0049-3848/82/230637-11$03.00/O Printed in the USA. Copyright (c> 1982 Pergamon Press Ltd. All rights reserved.
PROSTACYCLIN BIOSYNTHESIS IN VASCULAR ENOOTHELIUM IS NOT INHIBITED BY CYCLIC AMP. STUDIES WITH 3-ISOBUTYL-l-METHYLXANTHINE AND FORSKOLIN.
Abigail F. Adams Brotherton, Donald E. Macfarlane and John C. Hoak Cardiovascular Center and Division of Hematology-Oncology, Department of Medicine University of Iowa, Iowa City, Iowa 52242
(Received 30.6.1982; in revised form 30.8.1982. Accepted by Editor J.B. Smith. Received in final form by Executive Editorial Office 14.9.1982.
ABSTRACT
We have previously reported (Proc. Natl. Acad. Sci. 79, 495-499, 1982) that the cyclic nucleotide phosphodiesterase inhibitor, 3-isobutyl-1-methylxanthine (IBMX), stimulates cyclic AMP accumulation and inhibits prostacyclin (PGIz) production in primary monolayer cultures of human umbilical vein endothelium. The present study was carried out to determine whether these effects are causally related. Incubation of endothelial monolayers with the diterpene, forskolin, increased the intracellular concentration of cyclic AMP by lo-fold. Despite this marked increase in cyclic AMP, neither baseline production of PGI;!nor release in response to stimulation by thrombin or the divalent cation ionophore, A23187, was affected. Both forskolin and isoproterenol were found to potentiate the effect of IBMX on cyclic AMP accumulation without causing further inhibition of PGI;! biosynthesis. Inhibition of cyclic nucleotide phosphodiesterase activity with 2,6-bis-(diethanolamino)-4_piperidinopyrimido[5,4-ilpyrimidine increased cyclic AMP levels to the same extent as IBMX; however, this agent had no effect on PGI;!biosynthesis. These findings demonstrate that increases in the intracellular concentration of cyclic AMP have no short-term effects on PGI2 biosynthesis in vascular endothelium and suggest that inhibition of PGI2 production by IBMX is the result of some other, cyclic AMP-independent action of the drug.
Key Words:
prostacyclin, endothelium, cyclic AMP, methylxanthine, forskolin 637
IYTRODUCTIGN Prostacyciin (PGi;) is a potent vasodilatcr, inhibitor cf platelet aggregation (I) and agonist of adenylate cyclase (2) that is produced by the vascular endothelium in response to a variety of stimuli including thrombin and the divalent cation ionophore A23187 (3). We (4-6) and others (7) have reported that the cyclic nucleotide phosphodiesterase inhibitor, 3-isobutyll-methylxanthine (IBMX), stimulates cyclic AMP accumulation and inhibits PGI, biosynthesis in cultured umbilical vein endothelium. This observation suggested that the inhibitory effect of IBMX may be mediated by cyclic AMP and that cyclic AMP may be involved in the regulation of PGI, production by the vascular endothelium. The purpose of the present study was to determine whether inhibition of PGI, biosynthesis by IBM]! is due to its stimuiatory effect on cyclic AMP accumulation or whether it is the result of some other, cyclic AMP-independent action of the drug. Preliminary results have been presented in abstract form (8). MATERIALS
AND METHODS
Materials. 2,6-Bis(diethanolamino)-4,8-dipiperidinopyrimido-[5,4,d]pyrimmpyridamole) and L-isoproterenol hydrochloride were purchased from Sigma. Forskolin was obtained from Calbiochem-Behring. ["sI]Tyrosine methyl ester-2'-O-succinyl cyclic AMP (2000-3000 Ci/mmol), [2,8-aH]cyclic AMP (38.1 Ci/mmol), [l"C(U)]arachidonic acid (500 mCi/mmol); and, cyclic AMP antiserum complex were purchased from New England Nuclear. Z-6-Bis(diethanolamino)-4-piperidinopyrimido-[5,4ld]-pyrimidine (RA 233) was a gift from Dr. J.W. Bell of Boehringer Ingleheim, Ltd., Isleworth, Middx, UK. The sources of other supplies are described elsewhere (6, 9-12). Cell culture. Primary cultures of human endothelial cells from umbilical veins were prepared by a modification (9) of the method of Jaffe et al. (13). Confluent endothelial cell monolayers containing approximately 4.5 x lo5 cells per 24 mm diameter well of a 12-well culture plate (Linbro) were used 4 days after seeding. Incubation procedure. Immediately prior to each experiment, culture medium was aspirated and monolayers were rinsed twice with 1 ml of "H/H buffer" (Hank's balanced salt solution (without NaHCO,)(Gibco); buffered to pH 7.4 with 15 mM 4-(2-hydroxyethyl)-1-piperazine-ethane sulfonic acid (HEPES). Monolayers were preincubated at 37°C for 10 min with 450 p1 of H/H buffer alone or containing the indicated test reagent(s). At the end of the preincubation period, 50 ~1 of either H/H buffer or buffer containing thrombin or A23187 was added to the monolayers to give a total volume of 500 ~1 at the concentration indicated. Monolayers were incubated at 37°C for a further 5 min on a rocker platform. Incubation medium was removed from the monolayers and frozen at -20°C. Extraction and isolation of cyclic AMP. Cyclic AMP was extracted from endothelial monolayers as previously described (6). To isolate cyclic AMP, acid extracts were added to columns (Pasteur pipettes) containing 1 ml of packed Dowex AG 5OW-4X resin (Sigma) that had been washed with 0.05 N HCl. After addition of 4 ml of 0.05 N HCl, cyclic AMP was eluted with 2 ml of H,O and collected into culture tubes containing 400 L1 of 30% (wt/vol) trichloroacetic acid. Acidified eluates were added to columns (Pasteur pipettes) containing approximately 0.89 of neutral alumina that had been Columns were washed with 5 ml of H,O rinsed with 5% trichloroacetic acid.
followed by 1 ml of 0.2 M Na-acetate buffer, pH 6.2; and cyclic AMP-was Recovery of i'H]cyclic eluted with an additional 1 ml of Na-acetat- buffer. AMP ranged from 65 to 70;;. Radioimmunoassay of cyclic AMP. 100 :l of sample or standard solution (in 0.2 M Na-acetate buffer) was transferred to a micro test tube and 5 -1 of acetic anhydrideltriethylamine (1:Z) was added followed by immediate Assay tubes (10 x 75 mm disposable glass culture tubes) were vortex mixing. 40 ~1 of acetylated sample or standard; 40 ~1 of a prepared by adding: methyl ester-2'-Osolution containing 8000 to 12,000 cpm of [ "lI]tyrosine succinyl cyclic AMP and 1% (wt/vol) normal rabbit serum in 0.2 M Na-acetate buffer; and, 40 ;?l of cyclic AMP antiserum complex diluted according to Smaller sample volumes may be assayed provided manufacturers instructions. that acetylated 0.2 Cl Na-acetate buffer is added to give a total volume of 40 pl. Alternatively, samples may be diluted with 0.2 M Na-acetate buffer Eight concentrations of cyclic AMP were included in prior to acetylation. the standard curve which ranged from 2.5 to 500 fmol/40 ul. Under the above conditions, 50% displacement of ["51]tyrosine methyl ester-?'-O-succinyl cyclic AMP from cyclic AMP antibody is obtained with 20 to 22 fmol of cyclic AMP/40 ~1, and approximately 30% of average net total counts are bound in the absence of cyclic AMP. All samples were assayed at two dilutions and the concentration of cyclic AMP was corrected for recovery. The radioimmunoassay of 6-keto-PGF:, Radioimmunoassay of 6-keto-PGF,,. is described elsewhere (11). Assay detection limits are 0.25 pmol of 6-keto-PGF,, per ml, and 50% inhibition is obtained with 1.8 to 2.2 pmol of 6-keto-PGF,, per ml. This assay has 4% crossreactivity with PGF1,, 2% with PGF,,, 1.6% with PGE,, and 1% with PGE,, PGD or PGD2. Assay for release of [14C]arachidonate. Release of radiolabeled arachidonate from endothelial cell monolayers was measured by a modification of the method described by Lollar and Owen (14). Phospholipjd pools were prelabeled by adding 0.05 &i of [14C]arachidonic acid in ethanol (5 ~1) to the culture medium (1 ml) covering each monolayer. Monolayers were incubated with [14C]arachidonic acid for 6 to 8 h; approximately 80% of the l'+C became cell-associated during this time. Aspirin (5 ~1 of a 100 mM solution in ethanol) was added to monolayers 1 h prior to each experiment; this treatment inhibited PG12 release to c2.5 nM. Experiments were started by aspirating the culture medium and rinsing monolayers with 1 ml of H/H buffer containing 0.04% (wt/vol) fatty acid poor bovine serum albumin (Fraction V; Pentex). Incubations with test reagents were carried out as described above, with the exception that 0.04% bovine serum albumin was added to the H/H buffer. Release of l'+Cwas measured by liquid-scintillation counting of a 200 ~1 sample of incubation medium. Preliminary experiments, in which incubation medium from selected samples was extracted and chromatographed as described by Marcus et al. (15), showed that greater than 90% of the radioactivity released from aspirin-treated monolayers migrated with arachidonic acid. Protein determination. Protein was determfned by the method of Lowry et al. (16) using bovine serum albumin as standard. On average, the concentration of protein in acid-precipitated monolayers of endothelial cells from 40 separate pools was 136 f 6 gg of protein/4.5 x lo5 cells (mean ? SE).
640
701.28, X0.5
CYCLIC AXP AXD ?GI? BIOS~V'TIIESIS
RESULTS The relationship between cyclic AMP accumulation and PGI2 biosynthesis was re-examined by first determining whether a stimulus of adenylate cyclase, such as isoproterenol, potentiates the effect of IBMX in primary endothelium. In the absence of IBEX, 10 PM isoproterenol produced a small increase in cyclic AMP which had no effect on either thrombin or A23187 induced release of PG12 (Fig. 1). A similar increase in cyclic AMP in response to isoproterenol was seen in six separate experiments; in each case, this effect was determined to be statistically significant (p < 0.05) by the unpaired Student's t test. Preincubation of endothelial monolayers with 1 mM IBMX increased cyclic AMP levels by Z-fold and inhibited PGI, release by about 60% (Fig. 1). Although 10 PM isoproterenol potentiated the effect of 1 mM IBMX on cyclic AMP accumulation, which resulted in levels of cyclic AMP comparable to those obtained with 4 mM IBMX alone, no additional inhibition of PGI2 production was observed. Note that 4 mM IBMX inhibited PGI2 release in response to either stimulus by nearly 100%
&Keto-PCiFta(nM1
600 400 200 0 Control
Cyclic AMP (pmol/ 4.5 x 10*cells)
ISO. l*M
IBMX.
1mM
SO. 1OyM ?? IEHX.
1.61 f 0.05
1.88 f 0.03
3.19 f 0.10
IElMX.4mM
1mM
5.08 f 0.08
4.08 f 0.14
ISO. 1OvM + lBMX.4mM
6.67 f 0.18
FIG. 1 Effect of isoproterenol and IBMX on cyclic AMP accumulation and PGI, production. Incubations with additions at the concentrations ind'ic'ated were carried out as described under "Methods". Values given for cyclic AMP are from monolayers that were not exposed to thrombin or A23187 (@); these values reflect the concentration of cyclic AMP present in monolayers after preincubation and prior addition of either stimulus. Values given are means + SE of triplicate determinations from the same experiments. Similar findings were obtained in three other experiments. Isoproterenol is abbreviated to ISO.
v01.28,
641
CYCLTC AMI'A.NDPGZ2 B'IOSYXTIIESIS
No.5
To generate larger increases in cyclic AMP accumulation, endothelial Forskolin has monolayers were preincubated with the diterpene, forskolin. been reported to cause a rapid and reversible activation of adenylate cyclase by a mechanism that appears to involve a direct action on the catalytic subunit of the enzyme (17,18). As shown in Fig. 2, the intracellular concentration of cyclic AMP in monolayers preincubated with 10 FM-forskolin alone was increased by an average of about lo-fold; however, forskolin had no effect on either thrombin or A23187 induced release of PGI,. Addition of forskolin together with 0.4 mM IBMX increased cyclic AMP levels by about 28-fold as compared to 0.4 ~$1 IBMX alone. Despite the marked potentiation of cyclic AMP accumulation by forskolin, no further inhibition of PGI, release in response to stimulation by either thrombin or A23187 was observed.
6-Keto-PGFla(nM)
: :.:..
‘..
.. :
;i:,
.:,: .,. ..
.\:
Thrombin.
O.Sp/mi
:.I.’ 7. y
;s :...
0
:. .. :.. .. ... :..: .... ::.:. ...
.. ...
~~ contml
CyclicAMP 4.5 x(pmol/ 105celIs)
.:,. .;
J.54 0.09 f
._
[email protected]
15.13 0.74
ISMX.0 4mM
*
2.82 f 0.35
FSK. l@M + IBuX.O.4mM
77.67 9.71
f
IEMX.4mM
5.16 f 0.40
FSK. tothi ISMX.4mU
105.34 11.13
f
FIG. 2 Effect of forskolin and IBMX on cyclic AMP accumulation and PGI? production. Incubations with additions at the concentrations indicated were carried out as described under "Methods". Values given for cyclic AMP are from monolayers that were not exposed to thrombin or A23187 (a); these values reflect the concentration of cyclic AMP present in monolayers after preincubation and prior to addition of either stimulus. Values given are means 2 SE of triplicate determinations from the same experiment. Similar findings were obtained in four other experiments. Forskolin is abbreviated to FSK.
The capacity of forskolin, and to a lesser extent, isoproterenol to potentiate the effect of IBMX on cyclic AMP accumulation, but not on PGI2 production, suggests that inhibition of PGI2 release by IBMX is not causally related to its effect on the total intracellular concentration of cyclic AMP in endothelial monolayers. To test this hypothesis further, we examined the effects of other cyclic nucleotide phosphodiesterase inhibitors on both
cyclic AMP accumulation and PGIz release. Preincsaation of endothelial monolayers with 4 r&l IBMX, 40 rr&!theophylline, 100 ;M dipyridamole or 100 LF! RA 233 resulted in comparable increases in Lhe intracellular concentration of cyclic AMP (Fig. 3). However, complete inhibition of thrombin and A23187 induced release of PGI, was observed only with IBMX. At a near-saturating concentration of 40 ITIM,theophylline inhibited thrombin and A23187 induced In the presence of 100 tiiM release of PGIz by about 70% (Fig. 3). dipyridamole, thrombin or A23187 induced release of PGI- was decreased by 40 and 50? respectively (Fig. 3). By contrast, RA 233, which is a derivative of dipG&damole, had a small stimulatory effect on release in response to both stimuli (Fig. 3). In the experiment shown in Fig. 3 as in five other experiments, the stimulatory effect of RA 233 on PGI, release was determined to be statistically significant (p <0.05)-by the unpaired Student's -t test.
6-Keto-PGFlo(nM) 800 Thrombin.
?? A23187.
600
OSp/ml lOuP.
400
200
0
Control
(pmoi/ 4.5 x 105cells)
IBMX 4mM
Cyclic AMP 1.96 k 0.08
5.13 i 0.10
Thsoohylline. 40mM 4.44
0.18
i
Digyridamole.
RA 233.
lOOpI.
lDDVh4
4.20 f 0.07
4.77 f 0.10
FIG. 3 Comparison of the effects of IBMX, theophylline, dipyridamole and PA 233 on cyclic AMP accumulation and PGI2 production. Incubations with additions at the concentrations indicated were carried out as described under "Methods". Values given for cyclic AMP are from monolayers that were not exposed to thrombin or A23187 (a); these values reflect the concentration of cyclic AMP present in monolayers after preincubation and prior to addition of either stimulus. Values given are means ?:SE of triplicate determinations from the same experiment. Similar findings were obtained in two other experiments.
To determine whether IBMX, theophylline and dipyridamole act by inhibiting the metabolism of arachidonic acid we studied the effects of these agents on thrombin and A23187 induced release of [lLC]arachidonic acid Results of a previous study from our from aspirin-treated endothelium.
vo1.25,
No.5
CYCLIC AMP AND PGI2 BIOSYNTHESIS
643
laboratory (6) suggested that part of the inhibitory effect of IBMX may be attributable to direct inhibition of cyclooxygenase activity because 4 mM IBMX decreased arachidonic acid induced release of PGIi by an average of As shown in Table 1, IBMX 50%, but had no effect on PGHp induced release. inhibited thrombin or A23187 induced release of [!iC]arachidonic acid by E-70% whereas PGI? release was decreased by nearly 100%. This finding suggests that about one-third of the inhibitory effect of IBMX can be attributed to inhibition of cyclooxygenase activity and two-thirds to In contrast inhibition of an earlier step in the pathway of PGI, release. to IBMX, theophylline decreased both [lY]arachidonic acid and PGIz release by 50-60X, which suggests that theophylline may not affect cyclooxygenase activity. Dipyridamole had a relatively small effect on [l‘+C]arachidonic acid release (20%) compared to its effect on total PGI2 release (70%), which suggests that dipyridamole primarily acts by inhibiting cyclooxygenase and/or PGI, synthetase activities. RA 233 had no significant inhibitory effect on release of [l'+C]arachidonic acid or PGI2.
TABLE 1 Comparison of the effects of cyclic nucleotide phosphodiesterase inhibitors on release of [14C]arachidonic acid and PGI2 Incubations with additions at the concentrations indicated were carried out as described under "Methods" with the exception that all solutions were prepared with H/H buffer containing 0.04% fatty acid poor bovine serum albumin. Endothelial monolayerswere prelabeled with [14C]arachidonic acid and treated with aspirin as described under "Methods". Thrombin (0.5 u/ml) and A23187 (4 uM) induced the release of 3372 + 103 and 8067 + 215 cpm [14C]arachidonic acid from aspirin-treated monolayers. Release of PGI2 from untreated monolayers in response to thrombin or A23187 was 252 + 8 and 396 ? 18 nM-6-keto-PGFI,, respectively. Percent inhibition of release was calculated after subtraction of baseline release of eitherl?-cpm or nM-6-keto-PGFlo. Values given are means ?rSE of triplicate determinations. Similar findings were obtained in two other experiments.
Stimulus
Inhibitor
Inhibition of arachidonic acid release
Inhibition of PGI, release
% Thrombin
A23187
None IBMX (4 mM) Theophylline (40 mM) Dipyridamole (100 uM) RA 233 (100 @l)
or3 65 2 8 57 t 4 19 + 5 8 ?:6a
None IBMX (4 mM) Theophylline (40 mM) Dipyridamole (100 uM) RA 233 (100 uM)
023 67 + 2 61 2 2 23 + 2 0 +-3a
% 023
98 + 1
56 ?:4 73 + 4 4 + 2a 025 99
2
1
50 + 8 70 +_6 1 z ga
a Not significantly different from zero (p > 0.05 by unpaired Student's r test).
544
OIWSSION Stimuli of adenylate cyclase such as PGI, (4,6), adenosine (19) and isoproterenol (Fig. 1) have little or no effect on cyclic AMP accumulation in primary vascular endothelium in the absence of a cyclic nucleotide phosphodiesterase inhibitor. In initial studies of the role of cyclic AMP in the regulation of PG12 biosynthesis, the methylxanthine, IBMX, was used to inhibit cyclic AMP phosphodiesterase activity and thereby promote the accumulation of cyclic AMP. The finding that IBMX caused a dose-dependent increase in the intracellular concentration of cyclic AMP as well as a dose-dependent inhibition of PG12 biosynthesis led us (4-6) and Hopkins and German (7) to propose that these effects were causally related. However, results of the present study refute this hypothesis. Unlike other stimuli of adenylate cyclase, forskolin increased the intracellular concentration of cyclic AMP by lo-fold in the absence of IBMX. Despite this marked increase in cyclic AMP, neither baseline production of PGI2 nor release in response to thrombin or A23187 was affected (Fig. 2). This finding alone directly demonstrates that increases in the intracellular concentration of cyclic AMP have no short-term effects on PGI, biosynthesis and suggests that cyclic AMP does not mediate the inhibition of PGI, biosynthesis by IBMX. Further evidence in support of this hypothesis is provided by the observation that both forskolin and isoproterenol potentiate the stimulatory effect of IBMX on cyclic AMP accumulation but neither agent potentiates the inhibition of PG12 biosynthesis by IBMX (Figs. 1 and 2). By contrast, Hopkins and Gorman (7) reported that isoproterenol potentiated the effect of IBMX on cyclic AMP accumulation as well as the inhibition of thrombin-induced release of PG12. Because these workers used subcultured vascular endothelium and did not measure cyclic AMP accumulation and PG12 biosynthesis in the same monolayers, direct comparisons with our data cannot be made. Adler et al. (20) reported that isoproterenol alone inhibits A23187 induced release of arachidonate from aspirin-treated subcultured vascular endothelium and that this effect was "markedly enhanced" by the addition of 3 mM IBMX. This finding is uninterpretable as the effect of IBMX alone was not shown. In light of our results, the effect of isoproterenol on PG12 production and arachidonate release reported by Hopkins and Gorman (7) and by Adler et al. (20) respectively, are not likely to be mediated by cyclic AMP. The finding that RA 233 can elevate cyclic AMP levels to the same extent as IBMX without affecting PG12 biosynthesis or the release of arachidonic acid from aspirin-treated endothelium is further evidence that the inhibitory effect of IBMX is not mediated by cyclic AMP. RA 233 is a pyrimidopyrimidine compound that has been shown to inhibit cyclic AMP phosphodiesterase activity in human blood platelets (21). In platelets, RA 233 and its parent compound dipyridamole, were found to be almost equally effective in inhibiting cyclic AMP phosphodiesterase activity (21). Although equimolar concentrations of RA 233 and dipyridamole produced comparable increases in the intracellular concentration of cyclic AMP in This endothelial monolayersl only dipyridamole inhibited PGI, production. inhibitory effect of dipyridamole is clearly unrelated to its effect on cyclic AMP accumulation and thus is not a general property of all On the basis of our studies of arachidonate pyrimidopyrimidine compounds. release from aspirin treated endothelium, a major part of the inhibitory effect of dipyridamole can be attributed to direct inhibition of fatty acid This finding conflicts cyclooxygenase and/or PG12 synthetase activities.
Vo1.28, No.5
CYCLIC AM? AND PC+
BIOSY?ITHESTS
645
with a report by Blass et al. (22) that micromolar concentrations of dipyridamole stimulate the conversion of arachidonic acid and PGH, to PGI, in rat stomach fundus homogenates and microsomal fractions of pig aorta, respectively. The inhibition of PGI2 biosynthesis by IBMX is a complex phenomenon. To determine whether this effect of IBMX is a property shared by other methylxanthines, we studied the effect of theophylline on PGI, production. Although theophylline was also found to inhibit PGI, production, much higher concentrations were needed to produce an effect; maximal inhibition was observed with a near-saturating concentration of 40 mM. Interestingly, theophylline decreased arachidonic acid release and PGI, production by about the same extent. This finding suggests that theophylline inhibits PGIz. biosynthesis by a mechanism that does not involve the metabolism of By contrast, IBMX inhibited the release of arachidonic arachidonic acid. acid from aspirin-treated monolayers by about 65% and PGI2 production by In a previous study (6), we found that IBMXpartially inhibited nearly 100%. arachidonic acid induced release of PGI2 but had no effect on PGH:! induced release. Together, these findings suggest that only one-third of the inhibitory effect of IBMX can be attributed to inhibition of fatty acid cyclooxygenase activity. Thus, the remaining two-thirds of the inhibitory effect of IBMX and the entire inhibitory effect of theophylline may involve an earlier step in the pathway of release. At present, we can only speculate on the mechanism by which methylxanthines inhibit PGI2 biosynthesis in vascular endothelium. Since the induction of PG12 biosynthesis appears to be mediated by Ca2+ (6), we suggest that at millimolar concentrations, methylxanthines may block the mobilization and/or action of free cytosolic Ca 2+ by a mechanism distinct from wellcharacterized cyclic AMP-dependent mechanisms (23). Methylxanthines have recently been shown to have effects in other systems that are likewise unrelated to cyclic nucleotide accumulation (24); these include their ability to increase twitch tension in rat hemidiaphragms (25); to prolong the duration of systole (25); and, to inhibit bradykinin-induced synthesis of PGE2 in rabbit renal inner medullary slices (27). Although the mechanism of action of methylxanthines in these systems has not been elucidated, a decrease in the availability of free cytosolic Ca2+ could account for the observed effects. In light of the fact that increases in the intracellular concentration of cyclic AMP have no short-term effects on PGI, biosynthesis, the vascular endothelium may be an excellent model in which to study the effects of methylxanthines on the mobilization and action of Can+.
ACKNOWLEDGEMENTS We thank Patricia Riley for excellent technical assistance; Glenna Fry and Connie Schroeder for preparation of the cell cultures; and Wendy Marsh for typing the manuscript. This work was supported in part by Grants HL 14320-10 (Specialized Center of Research in Atherosclerosis), HL 22408-03, and HL-27561 from the National Heart, Lung and Blood Institute; and, Grant 81-852 from the American Heart Association.
REFERENCES 1.
MONCADA, S., HIGGS, E.A., and VANE, J.R. Human arterial and venous tissues generate prostacyclin (Prostaglandin X), a potent inhibitor of platelet aggregation. Lancet.i, 18-20, 1977.
2.
GORMAN, R.R. BUNTING, S., and MILLER, O.V. Modulation of human platelet adenylate cyclase by prostacyclin (PGX). Prostaglandins.13, 377-388, 1977.
3.
WEKSLER, B.B., LEY, C.W., and JAFFE, E.A. Stimulation of endothelial cell prostacyclin production by thrombin, tryspin, and ionophore J. Clin. Invest. g, 923-930, 1978. A23187.
4.
BROTHERTON, A.F.A., and HOAK, J.C. Role of Ca'+ ions and cyclic AMP in the regulation of PGI2 release from the vascular endothelium. Circulation.62, 165, 1980.
5.
HOAK, J.C., CZERVIONKE, R.L., FRY, G.L. HAYCRAFT, D.L., and BROTHERTON, A.F.A. Role of the vascular endothelium. Phi. Trans. R. Sot, Lond. 8294, 331-338, 1981.
6.
Role of Ca2' and cyclic AMP in the BROTHERTON, A.F.A., and HOAK, J.C. regulation of the production of prostacyclin by the vascular Proc. Natl. Acad. Sci. U.S.A. 79,495-499, 1982. endothelium.
7.
HOPKINS, N.K., and GORMAN, R.R. Regulation of endothelial cell cyclic nucleotide metabolism by prostacyclin. J. Clin. Invest. 67, 540-546, 1981.
8.
BROTHERTON, A.F.A., MACFARLANE, D.E., and HOAK, J.C. Cyclic AMP does not mediate the inhibition of PG12 release from the vascular endothelium by 3-isobutyl-1-methylxanthine. Fed. Proc. 5, 1305, 1982.
9.
SMITH, J.B., OGLETREE, M.L., LEFER, A.M., and NICOLAOU,.K.C. Antibodies which antagonize the effects of prostacyclin. Nature (Lond.) 274, 64-65, 1978.
10.
CZERVIONKE, R.L., HOAK, J.C., and FRY, G.L. Effect of aspirin on thrombin-induced adherence of platelets to cultured cells from the J. Clin. Invest. 62,847-856, 1979. blood vessel wall.
11.
CZERVIONKE, R.L., SMITH, J.B., HOAK, J.C., FRY, G.L., and HAYCRAFT, D.L. Use of radioimmunoassay to study thrombin-induced release of PGI, Thromb. Res. -781-786, 1979. from cultured endothelium.
12.
CZERVIONKE, R.L., SMITH, J.B., FRY, G.L., HOAK, J.C., and HAYCRAFT, Inhibition of prostacyclin by treatment of endothelium with D.L. J. Clin. Invest. 63, 1089-1092, 1979. aspirin.
13.
JAFFE, E.A., NACHMAN, R.L., BECKER, C.G., and MINICK, C.R. Culture of human endothelial cells derived from umbilical veins. Identification by morphologic and immunologic criteria. J. Clin. Invest. 52, 2745-2756, 1973.
Vol.23, No.5
CYCLIC Al%'.X-XI PGQ
6L7
BTOSYXTIESTS
14.
Evidence that the effects of thrombin on LOLLAR, P., and OWEN, W.G. arachidonate metabolism in cultured human endothelial cells are not mediated by a high affinity receptor. J. Biol. Chem. 255, 8031-8034, 1980.
15.
MARCUS, A.J., WEKSLER, B.B, and JAFFE, E.A. Enzymatic conversion of prostaglandin endoperoxide and arachidonic acid to prostacyclin by J. Biol. Chem. 253, 7138-7141, 1978. cultured human endothelial cells.
16.
LOWRY, O.H., ROSEBROUGH, N.J., FARR. A.L., and RANDALL, R.J. Protein measurement with the folin phenol reagent. J. Biol. Chem. 193, 265-275, 1951.
17.
SEAMON, K.B., PADGETT, W., and DALY, J.W. Forskolin: Unique diterpene activator of adenylate cyclase in membranes and in intact cells. Proc. Natl. Acad. Sci. U.S.A. 78, 3363-3367, 1981.
18.
SEAMON, K., and DALY, J.W. Activation of adenylate cyclase by the diterpene forskolin does not require the guanine nucleotide regulatory protein. J. Biol. Chem. -_, 256 9799-9801, i981.
19.
BROTHERTON, A.F.A., FRY, G.L., and HOAK, J.C. Comparison of the effects of PG12 and adenosine (ADO) on the concentration of cyclic AMP in cultured cells of vascular origin. -Blood. 58,Supp. 1, 231A, 1981.
20.
ADLER, B., GIMBRONE, M.A., Jr., SCHAFER, A.I., and HANDIN, R.I. Prostacyclin and 8-adrenergic catecholamines inhibit arachidonate release and PG12 synthesis by vascular endothelium. Blood.3 514-517, 1981.
21.
MILLS. D.C.B.. and SMITH. J.B. The influence on olatelet aaareaation of drugs that'affect the'accumulation of adenosine 3',5'- cycli; monophosphate in platelets. Biochem. J. 121, 185-196, 1971.
22.
BLASS, K.-E., BLOCK, H.-U., FQRSTER, W., and PONICKE, K. A potent stimulatory of prostacyclin (PGI2) biosynthesis. Pharmacol. 68, 71-73,1981.
23.
BERRIDGE, H.J. The interaction of cyclic nucleotides and calcium in the control of cellular activity. Adv. Cyclic Nucleotide Res. 6, l-98, 1975.
24.
WELLS, J.N., and KRAMER, G.L. Phosphodiesterase inhibitors as tools in cyclic nucleotide research: Mol. Cell. a precautionary comment. Endocrinol. 23, l-9, 1981.
25.
KRAMER, G.L., and WELLS, J.N. Xanthines and skeletal muscle: lack of relationship between phosphodiesterase inhibition and increased twitch tension in rat diaphragms. Mol. Pharmacol. 17, 73-78, 1980.
26.
MUSHLIN, P., BOERTH, R.C., and WELLS, J.N. Selective phosphodiesterase inhibition and alterations of cardiac function by alkylated xanthines. Mol. Pharmacol. 20, 179-189, 1981.
27.
Inhibition RAPP, N.S., ZENSER, T.V., MATTAMMAL, M.B., and DAVIS, B.B. of bradykinin stimulation of renal medullary prostaglandin E synthesis by phosphodiesterase inhibitors. J. Pharmacol. EXP. Ther. & 442-446, 1981.
Dipyridamo le: Br. J.
PROSTACYCLIN BIOSYNTHESIS IN VASCULAR ENOOTHELIUM IS NOT INHIBITED BY CYCLIC AMP. STUDIES WITH 3-ISOBUTYL-l-METHYLXANTHINE AND FORSKOLIN.
Abigail F. Adams Brotherton, Donald E. Macfarlane and John C. Hoak Cardiovascular Center and Division of Hematology-Oncology, Department of Medicine University of Iowa, Iowa City, Iowa 52242
(Received 30.6.1982; in revised form 30.8.1982. Accepted by Editor J.B. Smith. Received in final form by Executive Editorial Office 14.9.1982.
ABSTRACT
We have previously reported (Proc. Natl. Acad. Sci. 79, 495-499, 1982) that the cyclic nucleotide phosphodiesterase inhibitor, 3-isobutyl-1-methylxanthine (IBMX), stimulates cyclic AMP accumulation and inhibits prostacyclin (PGIz) production in primary monolayer cultures of human umbilical vein endothelium. The present study was carried out to determine whether these effects are causally related. Incubation of endothelial monolayers with the diterpene, forskolin, increased the intracellular concentration of cyclic AMP by lo-fold. Despite this marked increase in cyclic AMP, neither baseline production of PGI;!nor release in response to stimulation by thrombin or the divalent cation ionophore, A23187, was affected. Both forskolin and isoproterenol were found to potentiate the effect of IBMX on cyclic AMP accumulation without causing further inhibition of PGI;! biosynthesis. Inhibition of cyclic nucleotide phosphodiesterase activity with 2,6-bis-(diethanolamino)-4_piperidinopyrimido[5,4-ilpyrimidine increased cyclic AMP levels to the same extent as IBMX; however, this agent had no effect on PGI;!biosynthesis. These findings demonstrate that increases in the intracellular concentration of cyclic AMP have no short-term effects on PGI2 biosynthesis in vascular endothelium and suggest that inhibition of PGI2 production by IBMX is the result of some other, cyclic AMP-independent action of the drug.
Key Words:
prostacyclin, endothelium, cyclic AMP, methylxanthine, forskolin 637
IYTRODUCTIGN Prostacyciin (PGi;) is a potent vasodilatcr, inhibitor cf platelet aggregation (I) and agonist of adenylate cyclase (2) that is produced by the vascular endothelium in response to a variety of stimuli including thrombin and the divalent cation ionophore A23187 (3). We (4-6) and others (7) have reported that the cyclic nucleotide phosphodiesterase inhibitor, 3-isobutyll-methylxanthine (IBMX), stimulates cyclic AMP accumulation and inhibits PGI, biosynthesis in cultured umbilical vein endothelium. This observation suggested that the inhibitory effect of IBMX may be mediated by cyclic AMP and that cyclic AMP may be involved in the regulation of PGI, production by the vascular endothelium. The purpose of the present study was to determine whether inhibition of PGI, biosynthesis by IBM]! is due to its stimuiatory effect on cyclic AMP accumulation or whether it is the result of some other, cyclic AMP-independent action of the drug. Preliminary results have been presented in abstract form (8). MATERIALS
AND METHODS
Materials. 2,6-Bis(diethanolamino)-4,8-dipiperidinopyrimido-[5,4,d]pyrimmpyridamole) and L-isoproterenol hydrochloride were purchased from Sigma. Forskolin was obtained from Calbiochem-Behring. ["sI]Tyrosine methyl ester-2'-O-succinyl cyclic AMP (2000-3000 Ci/mmol), [2,8-aH]cyclic AMP (38.1 Ci/mmol), [l"C(U)]arachidonic acid (500 mCi/mmol); and, cyclic AMP antiserum complex were purchased from New England Nuclear. Z-6-Bis(diethanolamino)-4-piperidinopyrimido-[5,4ld]-pyrimidine (RA 233) was a gift from Dr. J.W. Bell of Boehringer Ingleheim, Ltd., Isleworth, Middx, UK. The sources of other supplies are described elsewhere (6, 9-12). Cell culture. Primary cultures of human endothelial cells from umbilical veins were prepared by a modification (9) of the method of Jaffe et al. (13). Confluent endothelial cell monolayers containing approximately 4.5 x lo5 cells per 24 mm diameter well of a 12-well culture plate (Linbro) were used 4 days after seeding. Incubation procedure. Immediately prior to each experiment, culture medium was aspirated and monolayers were rinsed twice with 1 ml of "H/H buffer" (Hank's balanced salt solution (without NaHCO,)(Gibco); buffered to pH 7.4 with 15 mM 4-(2-hydroxyethyl)-1-piperazine-ethane sulfonic acid (HEPES). Monolayers were preincubated at 37°C for 10 min with 450 p1 of H/H buffer alone or containing the indicated test reagent(s). At the end of the preincubation period, 50 ~1 of either H/H buffer or buffer containing thrombin or A23187 was added to the monolayers to give a total volume of 500 ~1 at the concentration indicated. Monolayers were incubated at 37°C for a further 5 min on a rocker platform. Incubation medium was removed from the monolayers and frozen at -20°C. Extraction and isolation of cyclic AMP. Cyclic AMP was extracted from endothelial monolayers as previously described (6). To isolate cyclic AMP, acid extracts were added to columns (Pasteur pipettes) containing 1 ml of packed Dowex AG 5OW-4X resin (Sigma) that had been washed with 0.05 N HCl. After addition of 4 ml of 0.05 N HCl, cyclic AMP was eluted with 2 ml of H,O and collected into culture tubes containing 400 L1 of 30% (wt/vol) trichloroacetic acid. Acidified eluates were added to columns (Pasteur pipettes) containing approximately 0.89 of neutral alumina that had been Columns were washed with 5 ml of H,O rinsed with 5% trichloroacetic acid.
followed by 1 ml of 0.2 M Na-acetate buffer, pH 6.2; and cyclic AMP-was Recovery of i'H]cyclic eluted with an additional 1 ml of Na-acetat- buffer. AMP ranged from 65 to 70;;. Radioimmunoassay of cyclic AMP. 100 :l of sample or standard solution (in 0.2 M Na-acetate buffer) was transferred to a micro test tube and 5 -1 of acetic anhydrideltriethylamine (1:Z) was added followed by immediate Assay tubes (10 x 75 mm disposable glass culture tubes) were vortex mixing. 40 ~1 of acetylated sample or standard; 40 ~1 of a prepared by adding: methyl ester-2'-Osolution containing 8000 to 12,000 cpm of [ "lI]tyrosine succinyl cyclic AMP and 1% (wt/vol) normal rabbit serum in 0.2 M Na-acetate buffer; and, 40 ;?l of cyclic AMP antiserum complex diluted according to Smaller sample volumes may be assayed provided manufacturers instructions. that acetylated 0.2 Cl Na-acetate buffer is added to give a total volume of 40 pl. Alternatively, samples may be diluted with 0.2 M Na-acetate buffer Eight concentrations of cyclic AMP were included in prior to acetylation. the standard curve which ranged from 2.5 to 500 fmol/40 ul. Under the above conditions, 50% displacement of ["51]tyrosine methyl ester-?'-O-succinyl cyclic AMP from cyclic AMP antibody is obtained with 20 to 22 fmol of cyclic AMP/40 ~1, and approximately 30% of average net total counts are bound in the absence of cyclic AMP. All samples were assayed at two dilutions and the concentration of cyclic AMP was corrected for recovery. The radioimmunoassay of 6-keto-PGF:, Radioimmunoassay of 6-keto-PGF,,. is described elsewhere (11). Assay detection limits are 0.25 pmol of 6-keto-PGF,, per ml, and 50% inhibition is obtained with 1.8 to 2.2 pmol of 6-keto-PGF,, per ml. This assay has 4% crossreactivity with PGF1,, 2% with PGF,,, 1.6% with PGE,, and 1% with PGE,, PGD or PGD2. Assay for release of [14C]arachidonate. Release of radiolabeled arachidonate from endothelial cell monolayers was measured by a modification of the method described by Lollar and Owen (14). Phospholipjd pools were prelabeled by adding 0.05 &i of [14C]arachidonic acid in ethanol (5 ~1) to the culture medium (1 ml) covering each monolayer. Monolayers were incubated with [14C]arachidonic acid for 6 to 8 h; approximately 80% of the l'+C became cell-associated during this time. Aspirin (5 ~1 of a 100 mM solution in ethanol) was added to monolayers 1 h prior to each experiment; this treatment inhibited PG12 release to c2.5 nM. Experiments were started by aspirating the culture medium and rinsing monolayers with 1 ml of H/H buffer containing 0.04% (wt/vol) fatty acid poor bovine serum albumin (Fraction V; Pentex). Incubations with test reagents were carried out as described above, with the exception that 0.04% bovine serum albumin was added to the H/H buffer. Release of l'+Cwas measured by liquid-scintillation counting of a 200 ~1 sample of incubation medium. Preliminary experiments, in which incubation medium from selected samples was extracted and chromatographed as described by Marcus et al. (15), showed that greater than 90% of the radioactivity released from aspirin-treated monolayers migrated with arachidonic acid. Protein determination. Protein was determfned by the method of Lowry et al. (16) using bovine serum albumin as standard. On average, the concentration of protein in acid-precipitated monolayers of endothelial cells from 40 separate pools was 136 f 6 gg of protein/4.5 x lo5 cells (mean ? SE).
640
701.28, X0.5
CYCLIC AXP AXD ?GI? BIOS~V'TIIESIS
RESULTS The relationship between cyclic AMP accumulation and PGI2 biosynthesis was re-examined by first determining whether a stimulus of adenylate cyclase, such as isoproterenol, potentiates the effect of IBMX in primary endothelium. In the absence of IBEX, 10 PM isoproterenol produced a small increase in cyclic AMP which had no effect on either thrombin or A23187 induced release of PG12 (Fig. 1). A similar increase in cyclic AMP in response to isoproterenol was seen in six separate experiments; in each case, this effect was determined to be statistically significant (p < 0.05) by the unpaired Student's t test. Preincubation of endothelial monolayers with 1 mM IBMX increased cyclic AMP levels by Z-fold and inhibited PGI, release by about 60% (Fig. 1). Although 10 PM isoproterenol potentiated the effect of 1 mM IBMX on cyclic AMP accumulation, which resulted in levels of cyclic AMP comparable to those obtained with 4 mM IBMX alone, no additional inhibition of PGI2 production was observed. Note that 4 mM IBMX inhibited PGI2 release in response to either stimulus by nearly 100%
&Keto-PCiFta(nM1
600 400 200 0 Control
Cyclic AMP (pmol/ 4.5 x 10*cells)
ISO. l*M
IBMX.
1mM
SO. 1OyM ?? IEHX.
1.61 f 0.05
1.88 f 0.03
3.19 f 0.10
IElMX.4mM
1mM
5.08 f 0.08
4.08 f 0.14
ISO. 1OvM + lBMX.4mM
6.67 f 0.18
FIG. 1 Effect of isoproterenol and IBMX on cyclic AMP accumulation and PGI, production. Incubations with additions at the concentrations ind'ic'ated were carried out as described under "Methods". Values given for cyclic AMP are from monolayers that were not exposed to thrombin or A23187 (@); these values reflect the concentration of cyclic AMP present in monolayers after preincubation and prior addition of either stimulus. Values given are means + SE of triplicate determinations from the same experiments. Similar findings were obtained in three other experiments. Isoproterenol is abbreviated to ISO.
v01.28,
641
CYCLTC AMI'A.NDPGZ2 B'IOSYXTIIESIS
No.5
To generate larger increases in cyclic AMP accumulation, endothelial Forskolin has monolayers were preincubated with the diterpene, forskolin. been reported to cause a rapid and reversible activation of adenylate cyclase by a mechanism that appears to involve a direct action on the catalytic subunit of the enzyme (17,18). As shown in Fig. 2, the intracellular concentration of cyclic AMP in monolayers preincubated with 10 FM-forskolin alone was increased by an average of about lo-fold; however, forskolin had no effect on either thrombin or A23187 induced release of PGI,. Addition of forskolin together with 0.4 mM IBMX increased cyclic AMP levels by about 28-fold as compared to 0.4 ~$1 IBMX alone. Despite the marked potentiation of cyclic AMP accumulation by forskolin, no further inhibition of PGI, release in response to stimulation by either thrombin or A23187 was observed.
6-Keto-PGFla(nM)
: :.:..
‘..
.. :
;i:,
.:,: .,. ..
.\:
Thrombin.
O.Sp/mi
:.I.’ 7. y
;s :...
0
:. .. :.. .. ... :..: .... ::.:. ...
.. ...
~~ contml
CyclicAMP 4.5 x(pmol/ 105celIs)
.:,. .;
J.54 0.09 f
._
[email protected]
15.13 0.74
ISMX.0 4mM
*
2.82 f 0.35
FSK. l@M + IBuX.O.4mM
77.67 9.71
f
IEMX.4mM
5.16 f 0.40
FSK. tothi ISMX.4mU
105.34 11.13
f
FIG. 2 Effect of forskolin and IBMX on cyclic AMP accumulation and PGI? production. Incubations with additions at the concentrations indicated were carried out as described under "Methods". Values given for cyclic AMP are from monolayers that were not exposed to thrombin or A23187 (a); these values reflect the concentration of cyclic AMP present in monolayers after preincubation and prior to addition of either stimulus. Values given are means 2 SE of triplicate determinations from the same experiment. Similar findings were obtained in four other experiments. Forskolin is abbreviated to FSK.
The capacity of forskolin, and to a lesser extent, isoproterenol to potentiate the effect of IBMX on cyclic AMP accumulation, but not on PGI2 production, suggests that inhibition of PGI2 release by IBMX is not causally related to its effect on the total intracellular concentration of cyclic AMP in endothelial monolayers. To test this hypothesis further, we examined the effects of other cyclic nucleotide phosphodiesterase inhibitors on both
cyclic AMP accumulation and PGIz release. Preincsaation of endothelial monolayers with 4 r&l IBMX, 40 rr&!theophylline, 100 ;M dipyridamole or 100 LF! RA 233 resulted in comparable increases in Lhe intracellular concentration of cyclic AMP (Fig. 3). However, complete inhibition of thrombin and A23187 induced release of PGI, was observed only with IBMX. At a near-saturating concentration of 40 ITIM,theophylline inhibited thrombin and A23187 induced In the presence of 100 tiiM release of PGIz by about 70% (Fig. 3). dipyridamole, thrombin or A23187 induced release of PGI- was decreased by 40 and 50? respectively (Fig. 3). By contrast, RA 233, which is a derivative of dipG&damole, had a small stimulatory effect on release in response to both stimuli (Fig. 3). In the experiment shown in Fig. 3 as in five other experiments, the stimulatory effect of RA 233 on PGI, release was determined to be statistically significant (p <0.05)-by the unpaired Student's -t test.
6-Keto-PGFlo(nM) 800 Thrombin.
?? A23187.
600
OSp/ml lOuP.
400
200
0
Control
(pmoi/ 4.5 x 105cells)
IBMX 4mM
Cyclic AMP 1.96 k 0.08
5.13 i 0.10
Thsoohylline. 40mM 4.44
0.18
i
Digyridamole.
RA 233.
lOOpI.
lDDVh4
4.20 f 0.07
4.77 f 0.10
FIG. 3 Comparison of the effects of IBMX, theophylline, dipyridamole and PA 233 on cyclic AMP accumulation and PGI2 production. Incubations with additions at the concentrations indicated were carried out as described under "Methods". Values given for cyclic AMP are from monolayers that were not exposed to thrombin or A23187 (a); these values reflect the concentration of cyclic AMP present in monolayers after preincubation and prior to addition of either stimulus. Values given are means ?:SE of triplicate determinations from the same experiment. Similar findings were obtained in two other experiments.
To determine whether IBMX, theophylline and dipyridamole act by inhibiting the metabolism of arachidonic acid we studied the effects of these agents on thrombin and A23187 induced release of [lLC]arachidonic acid Results of a previous study from our from aspirin-treated endothelium.
vo1.25,
No.5
CYCLIC AMP AND PGI2 BIOSYNTHESIS
643
laboratory (6) suggested that part of the inhibitory effect of IBMX may be attributable to direct inhibition of cyclooxygenase activity because 4 mM IBMX decreased arachidonic acid induced release of PGIi by an average of As shown in Table 1, IBMX 50%, but had no effect on PGHp induced release. inhibited thrombin or A23187 induced release of [!iC]arachidonic acid by E-70% whereas PGI? release was decreased by nearly 100%. This finding suggests that about one-third of the inhibitory effect of IBMX can be attributed to inhibition of cyclooxygenase activity and two-thirds to In contrast inhibition of an earlier step in the pathway of PGI, release. to IBMX, theophylline decreased both [lY]arachidonic acid and PGIz release by 50-60X, which suggests that theophylline may not affect cyclooxygenase activity. Dipyridamole had a relatively small effect on [l‘+C]arachidonic acid release (20%) compared to its effect on total PGI2 release (70%), which suggests that dipyridamole primarily acts by inhibiting cyclooxygenase and/or PGI, synthetase activities. RA 233 had no significant inhibitory effect on release of [l'+C]arachidonic acid or PGI2.
TABLE 1 Comparison of the effects of cyclic nucleotide phosphodiesterase inhibitors on release of [14C]arachidonic acid and PGI2 Incubations with additions at the concentrations indicated were carried out as described under "Methods" with the exception that all solutions were prepared with H/H buffer containing 0.04% fatty acid poor bovine serum albumin. Endothelial monolayerswere prelabeled with [14C]arachidonic acid and treated with aspirin as described under "Methods". Thrombin (0.5 u/ml) and A23187 (4 uM) induced the release of 3372 + 103 and 8067 + 215 cpm [14C]arachidonic acid from aspirin-treated monolayers. Release of PGI2 from untreated monolayers in response to thrombin or A23187 was 252 + 8 and 396 ? 18 nM-6-keto-PGFI,, respectively. Percent inhibition of release was calculated after subtraction of baseline release of eitherl?-cpm or nM-6-keto-PGFlo. Values given are means ?rSE of triplicate determinations. Similar findings were obtained in two other experiments.
Stimulus
Inhibitor
Inhibition of arachidonic acid release
Inhibition of PGI, release
% Thrombin
A23187
None IBMX (4 mM) Theophylline (40 mM) Dipyridamole (100 uM) RA 233 (100 @l)
or3 65 2 8 57 t 4 19 + 5 8 ?:6a
None IBMX (4 mM) Theophylline (40 mM) Dipyridamole (100 uM) RA 233 (100 uM)
023 67 + 2 61 2 2 23 + 2 0 +-3a
% 023
98 + 1
56 ?:4 73 + 4 4 + 2a 025 99
2
1
50 + 8 70 +_6 1 z ga
a Not significantly different from zero (p > 0.05 by unpaired Student's r test).
544
OIWSSION Stimuli of adenylate cyclase such as PGI, (4,6), adenosine (19) and isoproterenol (Fig. 1) have little or no effect on cyclic AMP accumulation in primary vascular endothelium in the absence of a cyclic nucleotide phosphodiesterase inhibitor. In initial studies of the role of cyclic AMP in the regulation of PG12 biosynthesis, the methylxanthine, IBMX, was used to inhibit cyclic AMP phosphodiesterase activity and thereby promote the accumulation of cyclic AMP. The finding that IBMX caused a dose-dependent increase in the intracellular concentration of cyclic AMP as well as a dose-dependent inhibition of PG12 biosynthesis led us (4-6) and Hopkins and German (7) to propose that these effects were causally related. However, results of the present study refute this hypothesis. Unlike other stimuli of adenylate cyclase, forskolin increased the intracellular concentration of cyclic AMP by lo-fold in the absence of IBMX. Despite this marked increase in cyclic AMP, neither baseline production of PGI2 nor release in response to thrombin or A23187 was affected (Fig. 2). This finding alone directly demonstrates that increases in the intracellular concentration of cyclic AMP have no short-term effects on PGI, biosynthesis and suggests that cyclic AMP does not mediate the inhibition of PGI, biosynthesis by IBMX. Further evidence in support of this hypothesis is provided by the observation that both forskolin and isoproterenol potentiate the stimulatory effect of IBMX on cyclic AMP accumulation but neither agent potentiates the inhibition of PG12 biosynthesis by IBMX (Figs. 1 and 2). By contrast, Hopkins and Gorman (7) reported that isoproterenol potentiated the effect of IBMX on cyclic AMP accumulation as well as the inhibition of thrombin-induced release of PG12. Because these workers used subcultured vascular endothelium and did not measure cyclic AMP accumulation and PG12 biosynthesis in the same monolayers, direct comparisons with our data cannot be made. Adler et al. (20) reported that isoproterenol alone inhibits A23187 induced release of arachidonate from aspirin-treated subcultured vascular endothelium and that this effect was "markedly enhanced" by the addition of 3 mM IBMX. This finding is uninterpretable as the effect of IBMX alone was not shown. In light of our results, the effect of isoproterenol on PG12 production and arachidonate release reported by Hopkins and Gorman (7) and by Adler et al. (20) respectively, are not likely to be mediated by cyclic AMP. The finding that RA 233 can elevate cyclic AMP levels to the same extent as IBMX without affecting PG12 biosynthesis or the release of arachidonic acid from aspirin-treated endothelium is further evidence that the inhibitory effect of IBMX is not mediated by cyclic AMP. RA 233 is a pyrimidopyrimidine compound that has been shown to inhibit cyclic AMP phosphodiesterase activity in human blood platelets (21). In platelets, RA 233 and its parent compound dipyridamole, were found to be almost equally effective in inhibiting cyclic AMP phosphodiesterase activity (21). Although equimolar concentrations of RA 233 and dipyridamole produced comparable increases in the intracellular concentration of cyclic AMP in This endothelial monolayersl only dipyridamole inhibited PGI, production. inhibitory effect of dipyridamole is clearly unrelated to its effect on cyclic AMP accumulation and thus is not a general property of all On the basis of our studies of arachidonate pyrimidopyrimidine compounds. release from aspirin treated endothelium, a major part of the inhibitory effect of dipyridamole can be attributed to direct inhibition of fatty acid This finding conflicts cyclooxygenase and/or PG12 synthetase activities.
Vo1.28, No.5
CYCLIC AM? AND PC+
BIOSY?ITHESTS
645
with a report by Blass et al. (22) that micromolar concentrations of dipyridamole stimulate the conversion of arachidonic acid and PGH, to PGI, in rat stomach fundus homogenates and microsomal fractions of pig aorta, respectively. The inhibition of PGI2 biosynthesis by IBMX is a complex phenomenon. To determine whether this effect of IBMX is a property shared by other methylxanthines, we studied the effect of theophylline on PGI, production. Although theophylline was also found to inhibit PGI, production, much higher concentrations were needed to produce an effect; maximal inhibition was observed with a near-saturating concentration of 40 mM. Interestingly, theophylline decreased arachidonic acid release and PGI, production by about the same extent. This finding suggests that theophylline inhibits PGIz. biosynthesis by a mechanism that does not involve the metabolism of By contrast, IBMX inhibited the release of arachidonic arachidonic acid. acid from aspirin-treated monolayers by about 65% and PGI2 production by In a previous study (6), we found that IBMXpartially inhibited nearly 100%. arachidonic acid induced release of PGI2 but had no effect on PGH:! induced release. Together, these findings suggest that only one-third of the inhibitory effect of IBMX can be attributed to inhibition of fatty acid cyclooxygenase activity. Thus, the remaining two-thirds of the inhibitory effect of IBMX and the entire inhibitory effect of theophylline may involve an earlier step in the pathway of release. At present, we can only speculate on the mechanism by which methylxanthines inhibit PGI2 biosynthesis in vascular endothelium. Since the induction of PG12 biosynthesis appears to be mediated by Ca2+ (6), we suggest that at millimolar concentrations, methylxanthines may block the mobilization and/or action of free cytosolic Ca 2+ by a mechanism distinct from wellcharacterized cyclic AMP-dependent mechanisms (23). Methylxanthines have recently been shown to have effects in other systems that are likewise unrelated to cyclic nucleotide accumulation (24); these include their ability to increase twitch tension in rat hemidiaphragms (25); to prolong the duration of systole (25); and, to inhibit bradykinin-induced synthesis of PGE2 in rabbit renal inner medullary slices (27). Although the mechanism of action of methylxanthines in these systems has not been elucidated, a decrease in the availability of free cytosolic Ca2+ could account for the observed effects. In light of the fact that increases in the intracellular concentration of cyclic AMP have no short-term effects on PGI, biosynthesis, the vascular endothelium may be an excellent model in which to study the effects of methylxanthines on the mobilization and action of Can+.
ACKNOWLEDGEMENTS We thank Patricia Riley for excellent technical assistance; Glenna Fry and Connie Schroeder for preparation of the cell cultures; and Wendy Marsh for typing the manuscript. This work was supported in part by Grants HL 14320-10 (Specialized Center of Research in Atherosclerosis), HL 22408-03, and HL-27561 from the National Heart, Lung and Blood Institute; and, Grant 81-852 from the American Heart Association.
REFERENCES 1.
MONCADA, S., HIGGS, E.A., and VANE, J.R. Human arterial and venous tissues generate prostacyclin (Prostaglandin X), a potent inhibitor of platelet aggregation. Lancet.i, 18-20, 1977.
2.
GORMAN, R.R. BUNTING, S., and MILLER, O.V. Modulation of human platelet adenylate cyclase by prostacyclin (PGX). Prostaglandins.13, 377-388, 1977.
3.
WEKSLER, B.B., LEY, C.W., and JAFFE, E.A. Stimulation of endothelial cell prostacyclin production by thrombin, tryspin, and ionophore J. Clin. Invest. g, 923-930, 1978. A23187.
4.
BROTHERTON, A.F.A., and HOAK, J.C. Role of Ca'+ ions and cyclic AMP in the regulation of PGI2 release from the vascular endothelium. Circulation.62, 165, 1980.
5.
HOAK, J.C., CZERVIONKE, R.L., FRY, G.L. HAYCRAFT, D.L., and BROTHERTON, A.F.A. Role of the vascular endothelium. Phi. Trans. R. Sot, Lond. 8294, 331-338, 1981.
6.
Role of Ca2' and cyclic AMP in the BROTHERTON, A.F.A., and HOAK, J.C. regulation of the production of prostacyclin by the vascular Proc. Natl. Acad. Sci. U.S.A. 79,495-499, 1982. endothelium.
7.
HOPKINS, N.K., and GORMAN, R.R. Regulation of endothelial cell cyclic nucleotide metabolism by prostacyclin. J. Clin. Invest. 67, 540-546, 1981.
8.
BROTHERTON, A.F.A., MACFARLANE, D.E., and HOAK, J.C. Cyclic AMP does not mediate the inhibition of PG12 release from the vascular endothelium by 3-isobutyl-1-methylxanthine. Fed. Proc. 5, 1305, 1982.
9.
SMITH, J.B., OGLETREE, M.L., LEFER, A.M., and NICOLAOU,.K.C. Antibodies which antagonize the effects of prostacyclin. Nature (Lond.) 274, 64-65, 1978.
10.
CZERVIONKE, R.L., HOAK, J.C., and FRY, G.L. Effect of aspirin on thrombin-induced adherence of platelets to cultured cells from the J. Clin. Invest. 62,847-856, 1979. blood vessel wall.
11.
CZERVIONKE, R.L., SMITH, J.B., HOAK, J.C., FRY, G.L., and HAYCRAFT, D.L. Use of radioimmunoassay to study thrombin-induced release of PGI, Thromb. Res. -781-786, 1979. from cultured endothelium.
12.
CZERVIONKE, R.L., SMITH, J.B., FRY, G.L., HOAK, J.C., and HAYCRAFT, Inhibition of prostacyclin by treatment of endothelium with D.L. J. Clin. Invest. 63, 1089-1092, 1979. aspirin.
13.
JAFFE, E.A., NACHMAN, R.L., BECKER, C.G., and MINICK, C.R. Culture of human endothelial cells derived from umbilical veins. Identification by morphologic and immunologic criteria. J. Clin. Invest. 52, 2745-2756, 1973.
Vol.23, No.5
CYCLIC Al%'.X-XI PGQ
6L7
BTOSYXTIESTS
14.
Evidence that the effects of thrombin on LOLLAR, P., and OWEN, W.G. arachidonate metabolism in cultured human endothelial cells are not mediated by a high affinity receptor. J. Biol. Chem. 255, 8031-8034, 1980.
15.
MARCUS, A.J., WEKSLER, B.B, and JAFFE, E.A. Enzymatic conversion of prostaglandin endoperoxide and arachidonic acid to prostacyclin by J. Biol. Chem. 253, 7138-7141, 1978. cultured human endothelial cells.
16.
LOWRY, O.H., ROSEBROUGH, N.J., FARR. A.L., and RANDALL, R.J. Protein measurement with the folin phenol reagent. J. Biol. Chem. 193, 265-275, 1951.
17.
SEAMON, K.B., PADGETT, W., and DALY, J.W. Forskolin: Unique diterpene activator of adenylate cyclase in membranes and in intact cells. Proc. Natl. Acad. Sci. U.S.A. 78, 3363-3367, 1981.
18.
SEAMON, K., and DALY, J.W. Activation of adenylate cyclase by the diterpene forskolin does not require the guanine nucleotide regulatory protein. J. Biol. Chem. -_, 256 9799-9801, i981.
19.
BROTHERTON, A.F.A., FRY, G.L., and HOAK, J.C. Comparison of the effects of PG12 and adenosine (ADO) on the concentration of cyclic AMP in cultured cells of vascular origin. -Blood. 58,Supp. 1, 231A, 1981.
20.
ADLER, B., GIMBRONE, M.A., Jr., SCHAFER, A.I., and HANDIN, R.I. Prostacyclin and 8-adrenergic catecholamines inhibit arachidonate release and PG12 synthesis by vascular endothelium. Blood.3 514-517, 1981.
21.
MILLS. D.C.B.. and SMITH. J.B. The influence on olatelet aaareaation of drugs that'affect the'accumulation of adenosine 3',5'- cycli; monophosphate in platelets. Biochem. J. 121, 185-196, 1971.
22.
BLASS, K.-E., BLOCK, H.-U., FQRSTER, W., and PONICKE, K. A potent stimulatory of prostacyclin (PGI2) biosynthesis. Pharmacol. 68, 71-73,1981.
23.
BERRIDGE, H.J. The interaction of cyclic nucleotides and calcium in the control of cellular activity. Adv. Cyclic Nucleotide Res. 6, l-98, 1975.
24.
WELLS, J.N., and KRAMER, G.L. Phosphodiesterase inhibitors as tools in cyclic nucleotide research: Mol. Cell. a precautionary comment. Endocrinol. 23, l-9, 1981.
25.
KRAMER, G.L., and WELLS, J.N. Xanthines and skeletal muscle: lack of relationship between phosphodiesterase inhibition and increased twitch tension in rat diaphragms. Mol. Pharmacol. 17, 73-78, 1980.
26.
MUSHLIN, P., BOERTH, R.C., and WELLS, J.N. Selective phosphodiesterase inhibition and alterations of cardiac function by alkylated xanthines. Mol. Pharmacol. 20, 179-189, 1981.
27.
Inhibition RAPP, N.S., ZENSER, T.V., MATTAMMAL, M.B., and DAVIS, B.B. of bradykinin stimulation of renal medullary prostaglandin E synthesis by phosphodiesterase inhibitors. J. Pharmacol. EXP. Ther. & 442-446, 1981.
Dipyridamo le: Br. J.