Regulation of macrophage-mediated tumor cell killing by prostaglandins: comparison of the effects of PGE2 and PGI2

Mouse resident peritoneal macrophages stimulated in vitro by purified bacterial lipopolysaccharide (LPS) produced both prostaglandin E2 (PGE2) and prostaglandin I2 (PGI2), the latter detected as its stable metabolite, 6-keto PGF1 alpha. Maximum production, induced in each case by 1 ng/ml purified LPS, was in the range of 10(-7)M for PGI2 and 3 x 10(-8)M for PGE2. A quantitatively similar increase in intracellular levels of macrophage cyclic AMP (measured on a whole cell basis), with a similar duration of effect, was stimulated by PGE2 and PGI2; however, only PGE2 had a negative regulatory effect on macrophage activation for tumor cell killing. These data confirm that more than a whole cell increase in the concentration of cyclic AMP is needed to shut off nonspecific tumor cell killing mediated by LPS-activated resident peritoneal macrophages.     

A comparison of the effects of PGI2, PGE2 and PGH2 on the cyclic nucleotide levels in rat anterior pituitary gland in vitro.

Rat anterior pituitary explants were incubated with PGI2, PGH2 and PGE2 in the presence of theophylline (1mM) and the production of cyclic AMP was measured. PGE2 was found to be about 20 times more potent than PGI2 while PGH2 was slightly more effective than PGI2. The results suggest that PGI2 does not play a AMP was measured. PGE was found to be about 20 times more potent than physiological role in cyclic AMP mediated events in the rat anterior pituitary.     

Prostacyclin stimulation of dog arterial cyclic AMP levels.

Prostacyclin (PGI2) dose-dependently increases the adenosine 3',5'-cyclic monophosphate (cyclic AMP) levels in canine femoral, carotid, and canine and bovine coronary arteries. The prostacyclin-stimulation is enhanced by phosphodiesterase inhibitors, and is readily measurable after 60 sec incubation. The prostaglandin endoperoxide PGH2, but not PGH1, also elevates cAMP levels in femoral arteries. Inhibition of arterial prostacyclin synthetase with 28 microM 9,11-azoprosta-5,13-dienoic acid (azo analog I) blocks the PGH2-stimulation of cAMP accumulation. Azo analog I does not attenuate a direct PGI2 stimulation, indicating that the PGH2 dependent elevation of cAMP is due to conversion of PGH2 to PGI2 by the artery. PGI2 and PGE1 increase cyclic AMP levels and relax dog femoral and bovine coronary arteries, while PGE2, which actually contracts bovine coronary arteries, has no effect on arterial cyclic AMP levels. The significance of the PGI2-stimulation of arterial cyclic AMP is not known, but it is probably related to relaxation of arterial strips.     

Effects of prostaglandin E1, I2 and isoproterenol on the tissue cyclic AMP content in longitudinal muscle of rabbit intestine

Effects of prostaglandin E1(PGE1) and prostaglandin I2(PGI2) on the mechanical activity and tissue cyclic AMP content of the longitudinal muscle of rabbit intestine were examined, comparing that of the tissue cyclic AMP content. Isoproterenol caused a relaxation and increased tissue cyclic AMP content.     

The effects of hormones on cyclic adenosine 3':5'-monophosphate accumulation in transitional epithelium of the urinary bladder.

Transitional epithelium lining rabbit urinary bladders was isolated and studied in vitro. The homogeneity of the isolated epithelium was demonstrated by light and electron microscopical monitoring as well as cell culture studies. Transitional epithelium responded to epinephrine and prostaglandin E1 (PGE1) in the presence of 2mM 1-methyl, 3-isobutylxanthine (MIX) with increases in intracellular levels of cyclic adenosine 3':5'-monophosphate (cyclic AMP). Corticotropin, aldosterone, insulin, parathyroid hormone and vasopressin were slightly but significantly stimulatory under similar conditions. Glucagon and oxytocin were not stimulatory at the concentrations tested. The effects of epinephrine and PGE1 were potentiated by 2mM MIX 20-fold or greater. The cells were slightly more sensitive to PGE1 then to epinephrine. The prostaglandin produced a noticeable response at about 10nM, while effects of epinephrine were discernible at 0.1muM. Maximal responses to both effectors were seen at about 10muM. The action of 10muM epinephrine, but not 10muM PGE1, was completely abolished by 0.1mM propranolol. Responses to combinations of epinephrine and PGE1 were additive. Cyclic AMP accumulated in the incubation medium of transitional epithelial cells exposed to epinephrine, PGE1, MIX, or combinations of the agonists. The appearance of cyclic AMP in the medium was slow compared to the rate of intracellular accumulation, but reached significant levels following prolonged stimulation.     

Prostaglandin I2 stimulation of granulosa cell cyclic AMP production.

The ability of prostaglandin I2 (PGI2) to stimulate cyclic AMP production by granulosa cells, isolated from intact immature rats, has been demonstrated in vitro. The minimal effective dose was 15 ng/ml, which was comparable to the minimal effective dose for PGE2. However, a concentration of 15 microgram/ml PGI2 was required to stimulate cyclic AMP production maximally, compared to a concentration of 1 microgram/ml PGE2, which produced the maximum response. It therefore appears that PGI2 is not more effective than PGE2 in stimulating cyclic AMP production in granulosa cells, and is possibly less effective. Submaximal concentrations of PGI2 appeared to be able to modify the stimulation of cyclic AMP production by follicle-stimulating hormone (FSH), but whether or not PGI2 plays any role in follicular function remains to be established.     

Prostaglandin E2 inhibition of growth in a colony-stimulating factor 1-dependent macrophage cell line

The effects of prostaglandin E2 (PGE2) were examined in a murine macrophage cell line (BAC1.2F5) that was completely dependent on colony-stimulating factor-1 (CSF-1) for both growth and survival. The addition of PGE2 to cultures of BAC1.2F5 cells resulted in the inhibition of CSF-1-induced [3H]thymidine incorporation and cell proliferation. The inhibitory effects of PGE2 were mimicked by the addition of dibutyryl-cyclic AMP, and the effectiveness of PGE2 was markedly potentiated by 1-methyl-3-isobutylxanthine, a potent inhibitor of cyclic nucleotide phosphodiesterase activity. PGE2 caused a 10-fold elevation of the intracellular cyclic AMP concentration, whereas CSF-1 neither increased cyclic AMP levels nor attenuated the rise in cyclic AMP promoted by PGE2. However, CSF-1 may indirectly regulate cyclic AMP levels since in the absence of CSF-1, BAC1.2F5 cells actively synthesized PGE2, whereas PGE2 production was abruptly terminated by the addition of CSF-1. In BAC1.2F5 cells, PGE2 increases the intracellular cyclic AMP concentration, thereby blocking cell proliferation, but does not down-regulate the CSF-1 receptor or abrogate the functions of CSF-1 necessary for cell survival.     

Factors regulating the production of prostaglandin E2 and prostacyclin (prostaglandin I2) in rat and human adipocytes.

The regulation of PGE2 (prostaglandin E2) and PGI2 (prostaglandin I2; prostacyclin) formation was investigated in isolated adipocytes. The formation of both PGs was stimulated by various lipolytic agents such as isoproterenol, adrenaline and dibutyryl cyclic AMP. During maximal stimulation the production of PGE2 and PGI2 (measured as 6-oxo-PGF1 alpha) was 0.51 +/- 0.04 and 1.21 +/- 0.09 ng/2 h per 10(6) cells respectively. Thus PGI2 was produced in excess of PGE2 in rat adipocytes. The production of the PGs was inhibited by indomethacin and acetylsalicylic acid in a concentration-dependent manner. The half-maximal effective concentration of indomethacin was 328 +/- 38 nM and that of acetylsalicylic acid was 38.5 +/- 5.3 microM. The PGs were maximally inhibited by 70-75% after incubation for 2 h. In contrast with their effect on PG production, the two agents had a small potentiating effect on the stimulated lipolysis (P less than 0.05). The phospholipase inhibitors mepacrine and chloroquine inhibited both PG production and triacylglycerol lipolysis and were therefore unable to indicate whether the PG precursor, arachidonic acid, originates from phospholipids or triacylglycerols in adipocytes. Angiotensin II significantly (P less than 0.05) stimulated both PGE2 and PGI2 production in rat adipocytes without affecting triacylglycerol lipolysis. Finally, it was shown that PGE2 and PGI2 were also produced in human adipocytes, although in smaller quantities than in rat adipocytes. It is concluded that the production of PGs in isolated adipocytes is regulated by various hormones. Moreover, at least two separate mechanisms for PG production may exist in adipocytes: (1) a mechanism that is activated concomitantly with triacylglycerol lipolysis (and cyclic AMP) and (2) an angiotensin II-sensitive, but lipolysis (and cyclic AMP)-independent mechanism.     

Cyclic AMP does not inhibit A23187-induced prostaglandin biosynthesis in cultured rabbit renal microvascular endothelial cells.

We have previously shown that cultured rabbit renal preglomerular microvascular endothelial cells have the ability to synthesize a number of common prostaglandins. In the present study we have examined whether endogenous cyclic AMP is involved in the regulation of PGI2 and PGE2 biosynthesis in these cultured cells. Isoproterenol and forskolin produced an increase in cyclic AMP accumulation in these cells but had no effect on PGI2 or PGE2 biosynthesis either in the presence or absence of A23187. Similar results were noted in the presence of 3-isobutyl-1-methylxanthine, a cyclic AMP-phosphodiesterase inhibitor. These studies suggested that endogenous cyclic AMP does not regulate the biosynthesis of PGI2 or PGE2 in cultured renal preglomerular microvascular endothelial cells either under basal or A23187-stimulated condition. They further suggested that the effect of 3-isobutyl-1-methylxanthine on prostaglandin biosynthesis in these cultured cells was not secondary to its effects on phosphodiesterase.     

Inhibition of adenylate cyclase in human blood platelets by 9-substituted adenine derivatives.

A series of 9-substituted adenine derivatives inhibited adenylate cyclase activity (ATP pyrophosphate-lyase (cyclizing) EC 4.6.1.1) of a particulate preparation of human blood platelets. A 3--6 fold elevation of adenylate cyclase activity by prostaglandin E1 (PGE1) was inhibited in a concentration-related manner by 9-(tetrahydro-5-methyl-2-furyl) adenine (SQ 22,538), 9-(tetrahydro-2-furyl) adenine (SQ 22,536), 9-cyclopentyladenine (SQ 22,534), 9-furfuryladenine (sQ 4647) and 9-benzyladenine (SQ 218611). The I50 values ranged from 21 microM for SQ 22,538 to 140 microM for SQ 21,611. These same adenine derivatives reversed the inhibition by PGE1 of ADP-induced aggregation and the PGE1-stimulated elevation of adenosine 3':5'-monophosphate (cyclic AMP). The reversal of platelet aggregation inhibition by SQ 22,536 and SQ 4647 was concentration-related with I50 values of 30 microM in each case, whereas SQ 22,534 and SQ 21,611 reversed inhibition by 30% at 100 microM. SQ 22,536, SQ 22,534 and SQ 21,611 also blocked the increase in cyclic AMP levels in a concentration-related manner with I50 values of 1, 4 and 60 microM, respectively. SQ 4647 inhibited the elevation of cyclic AMP by more than 85% at 1000 microM. The adenine derivatives had no effect on platelet aggregation or on cyclic AMP levels in the absence of PGE1. These results provide additional evidence that the inhibition of platelet aggregation by PGE1 is mediated by cyclic AMP.     

PGE1-independent MDCK cells have elevated intracellular cyclic AMP but retain the growth stimulatory effects of glucagon and epidermal growth factor in serum-free medium

Prostaglandin E1 (PGE1), a component in the hormone-supplemented, serum-free medium for the Madin Darby canine kidney (MDCK) cell line, has been proposed to increase MDCK cell growth by increasing intracellular cyclic AMP levels. The association between increased intracellular cyclic AMP and the growth stimulatory effect of PGE1 has been examined in normal MDCK cells and in PGE1-independent variants of MDCK. These variant cells have lost the PGE1 requirement for long term growth in defined medium. Normal MDCK cells had almost twofold higher intracellular cyclic AMP levels during growth in Medium K-1 (9.0 pmol/mg protein) than in Medium K-1 minus PGE1. Furthermore, PGE1-independent clone 1 had higher intracellular cyclic AMP levels in Medium K-1 minus PGE1 than normal MDCK cells in Medium K-1. This latter observation suggests that the PGE1 requirement for MDCK cell growth is associated with the low intracellular cyclic AMP levels of this cell line. An involvement of cyclic AMP in the growth response to PGE1 is supported by these observations, as well as by the growth stimulatory effects of other agents that affect cyclic AMP metabolism in MDCK cells. These agents include glucagon, isobutyl methylxanthine (IBMX), and dibutyryl cyclic AMP. The growth of PGE1-independent clone 1 was inhibited rather than stimulated by PGE1. Similarly, PGE1-independent cell growth was inhibited by IBMX and dibutyryl cyclic AMP. However, the growth response to one agent which increases cyclic AMP (glucagon) was retained in PGE1-independent clone 1. This result suggests that the effect of glucagon is not associated with increases in intracellular cyclic AMP. The growth stimulatory effect of epidermal growth factor (EGF) on normal MDCK cells was also studied. Although EGF does not act via a cyclic AMP-mediated mechanism, EGF increased normal MDCK cell growth and substituted for PGE1 in Medium K-1. Thus, EGF and PGE1 could possibly affect similar growth-related functions in MDCK cells, although by different pathways. This possibility was examined further, using PGE1-independent clone 1. EGF, like glucagon, was still growth stimulatory to the PGE1-independent cells. Consequently, the biochemical pathways by which EGF and PGE1 increase MDCK cell growth probably do not converge.     

Calcium modulates the generation of inositol 1,3,4-trisphosphate in human platelets by the activation of inositol 1,4,5-trisphosphate 3-kinase.

Histamine (0.5 mM) stimulated the cyclic AMP content of cell suspensions containing greater than 80% parietal cells. Epidermal growth factor (EGF) inhibited this stimulatory effect of histamine, but had no effect on basal cyclic AMP content. The half-maximally effective concentration of EGF for inhibition of histamine-stimulated cyclic AMP was 3.9 nM. The equivalent measurement for the inhibition of histamine-stimulated aminopyrine accumulation was 3.0 nM. Aminopyrine accumulation was measured because it provides an index of the secretory activity of the cell. The cyclic AMP phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX) prevented the inhibitory effect of EGF on cyclic AMP content. This effect of IBMX was not caused by its ability to raise cellular cyclic AMP content in the presence of histamine. Prevention by IBMX of the inhibitory action of EGF on histamine-stimulated aminopyrine accumulation had been shown previously [Shaw, Hatt, Anderson & Hanson (1987) Biochem. J. 244, 699-704]. EGF stimulated prostaglandin E2 (PGE2) production in the cell fraction containing greater than 80% parietal cells, with the half-maximally effective concentration being 7.5 nM. EGF was ineffective in stimulating PGE2 production if the cell fraction was depleted of parietal cells (12%), or if 0.5 mM-histamine was added to the enriched parietal-cell fraction. In conclusion, EGF may inhibit histamine-stimulated acid secretion by decreasing the cyclic AMP content of parietal cells. This effect could be mediated by an increase in cyclic AMP phosphodiesterase activity, but it is unlikely to involve an effect of EGF on parietal-cell prostaglandin production.     

Prostaglandin biosynthesis and stimulation of cyclic AMP in primary monolayer cultures of epithelial cells from mouse mammary gland.

Cyclic AMP levels in primary monolayer cultures of epithelial cells prepared from mid-pregnant mice are stimulated by prostaglandin E1 and E2. Prostaglandin F1alpha and F2alpha have only a slight effect upon cyclic AMP levels. In the absence of phosphodiesterase inhibitors the rise in cyclic AMP produced by PGE1 is only transient and the levels return to normal within 30 minutes. High concentrations (16 mM) of theophylline are needed to prevent this decline, suggesting that the phosphodiesterase activity of epithelial cells in culture is high. However, theophylline alone produced only a small increase in basal cyclic AMP levels even over a 2-hour period indicating that basal cyclic AMP is turned over more slowly than cyclic AMP produced in response to stimulation with PGE1. Both PGE and PGF synthesis were monitored using radioimmunoassay procedures previously reported. The observed levels were found to decrease as cell density increased and were sensitive to the addition of agents such as collagen and naproxen.     

Action of luteinizing hormone-releasing hormone and synthesis of prostaglandin in the pituitary gland.

The possibility that prostaglandin E2 (PGE2) may play a role in luteinizing hormone (LH) release was examined using an in vitro model. Addition of luteinizing hormone-releasing hormone (LH-RH) to the culture medium stimulated cyclic AMP accumulation and LH-release by incubated hemipituitaries, but did not affect the level of PGE2 or prostaglandin synthetase activity in the gland. Aspirin and indomethacin reduced both prostaglandin synthetase activity and PGE2 or prostaglandin synthetase activity in the gland. Aspirin and indomethacin reduced both prostaglandin synthetase activity and PGE2 content in the pituitary, but did not impair the stimulatory action of LH-RH on either cyclic AMP accumulation or LH-release. Flufenamic acid on its own caused LH-release, but the drug abolished the effect of LH-RH on cyclic AMP accumulation. The mechanism of this action of flufenamic acid is not understood. It is concluded that the stimulatory action of LH-RH on pituitary cyclic AMP production and LH release is not mediated by prostaglandins.     

Effects of prostaglandin E1, isoproterenol and forskolin on cyclic AMP levels and tension in rabbit aortic rings

The role of cyclic AMP in the control of vascular smooth muscle tone was studied by monitoring the effects of prostaglandin E1 (PGE1), isoproterenol and forskolin on cyclic AMP levels and tension in rabbit aortic rings. PGE1, isoproterenol and forskolin all increased cyclic AMP levels in rabbit aortic rings. Isoproterenol and forskolin relaxed phenylephrine-contracted aortic rings, but PGE1 contracted the rings in the presence or absence of phenylephrine. Isoproterenol relaxed these PGE1-contracted aortic rings without further change in total cyclic AMP levels, which were already elevated by the PGE1 alone. Pretreatment with forskolin potentiated the effects of PGE1 on cyclic AMP levels. PGE1 caused contractions in muscles partially relaxed by forskolin, even though very large increases in cyclic AMP levels (30 fold) were produced by PGE1 in the presence of forskolin. Isoproterenol was able to relax these forskolin-treated, PGE1-contracted muscles with no further increase in cyclic AMP levels. Thus, there does not appear to be a good correlation between total tissue levels of cyclic AMP and tension in these experiments. Our results suggest that, if cyclic AMP is responsible for relaxation of smooth muscle, some form of functional compartmentalization of cyclic AMP must exist in this tissue.     

PGE1 and prostacyclin suppression of NK-cell mediated cytotoxicity and its relation to cyclic AMP

The capacity of three prostanoids (PGE1, 6-beta-PGI1, PGI2 or prostacyclin) and a phosphodiesterase inhibitor (rolipram) to inhibit NK ("natural killer") cell cytotoxicity and to raise cyclic AMP levels in purified NK cells was compared. PGE1 was about 200 times more potent than prostacyclin both in its ability to raise cyclic AMP and to inhibit NK cell cytotoxicity. The stable prostacyclin analogue, 6-beta-PGI1, had an intermediate potency. A 50% inhibition of cytotoxicity was obtained at approximately 10(-8) M for PGE1, 10(-7) M for 6-beta-PGI1, and 10(-6) M for both prostacyclin and rolipram. These doses raised the level of cyclic AMP by approximately 100%. These results suggest that PGE1 is likely to be more important as an endogenous regulator of lymphocyte cytotoxicity than prostacyclin. The results also provide further evidence that cyclic AMP is the mediator of prostanoid-induced reduction in NK cell activity.     

Endogenous synthesis of prostaglandins E2 and I2 in 3T3 fibroblasts transformed by polyoma virus: Effects on adenosine 3′:5′-monophosphate production

Prostaglandins (PG)E₁, E₂ and I₂ were produced by polyoma virus transformed (py) 3T3 fibroblasts. The levels of PGE₁, PGE₂ and 6-keto-PGF_1α (degradation product of PGI₂) were 22.7, 225 and 33.2 ng/ml medium, respectively, 72 h after medium change. The stimulatory potencies of exogenous PGE₁, PGE₂ and PGI₂ on adenosine 3′:5′-monophosphate (cyclic AMP) formation were similar. Therefore, the prostaglandin mediated increase in cyclic AMP levels observed during growth of these cells (Claesson, H.-E., Lindgren, J.Å. and Hammarström, S. (1977) Eur. J. Biochem. , 13) is largely (>80%) mediated by PGE₂ and to lesser extents by PGE₁ and PGI₂.     

Multiple signal transduction pathways lead to extracellular ATP-stimulated mitogenesis in mammalian cells: II. A pathway involving arachidonic acid release, prostaglandin synthesis, and cyclic AMP accumulation.

We have previously shown that extracellular ATP acts as a mitogen via protein kinase C (PKC)-dependent and independent pathways (Wang, D., Huang, N., Gonzalez, F.A., and Heppel, L.A. Multiple signal transduction pathways lead to extracellular ATP-stimulated mitogenesis in mammalian cells. I. Involvement of protein kinase C-dependent and independent pathways in the mitogenic response of mammalian cells to extracellular ATP. J. Cell. Physiol., 1991). The present aim was to determine if metabolism of arachidonic acid, resulting in prostaglandin E2 (PGE2) synthesis and elevation of cAMP levels, plays a role in mitogenesis mediated by extracellular ATP. Addition of ATP caused a marked enhancement of cyclic AMP accumulation in 3T3, 3T6, and A431 cells. Aminophylline, an antagonist of the adenosine A2 receptor, had no effect on the accumulation of cyclic AMP elicited by ATP, while it inhibited the action of adenosine. The accumulation of cyclic AMP was concentration dependent, which corresponds to the stimulation of DNA synthesis by ATP. The maximal accumulation was achieved after 45 min, with an initial delay period of about 15 min. That the activation of arachidonic acid metabolism contributed to cyclic AMP accumulation and mitogenesis stimulated by ATP in 3T3, 3T6, and A431 cells was supported by the following observations: (a) extracellular ATP stimulated the release of [3H]arachidonic acid and PGE2 into the medium; (b) inhibition of arachidonic acid release by inhibitors of phospholipase A2 blocked PGE2 production, cyclic AMP accumulation, and DNA synthesis activated by ATP, and this inhibition could be reversed by adding exogenous arachidonic acid; (c) cyclooxygenase inhibitors, such as indomethacin and aspirin, diminished the release of PGE2 and blocked cyclic AMP accumulation as well as [3H]thymidine incorporation in response to ATP; (d) PGE2 was able to restore [3H]thymidine incorporation when added together with ATP in the presence of cyclooxygenase inhibitors; (e) pertussis toxin inhibited ATP-stimulated DNA synthesis in a time- and dose-dependent fashion as well as arachidonic acid release and PGE2 formation. Other evidence for involvement of a pertussis toxin-sensitive G protein(s) in ATP-stimulated DNA synthesis as well as in arachidonic acid release is presented. In A431 cells, the enhancement of arachidonic acid and cyclic AMP accumulation by ATP was partially blocked by PKC down-regulation, implying that the activation of PKC may represent an additional pathway in ATP-stimulated metabolism of arachidonic acid. In all of these studies, ADP and AMP-PNP, but not adenosine, were as active as ATP.(ABSTRACT TRUNCATED AT 400 WORDS)     

Effects of protein kinase C activation on human platelet cyclic AMP metabolism

Treatment of intact human platelets with the tumour-promoting phorbol ester, phorbol 12-myristate 13-acetate (PMA), specifically inhibited PGD2-induced cyclic AMP formation without affecting the regulation of cyclic AMP metabolism by PGI2, PGE1, 6-keto-PGE1, adenosine or adrenaline. This action of PMA was: (i) concentration-dependent; (ii) not mediated by evoked formation or release of endogenous regulators of adenylate cyclase activity (thromboxane A2 or ADP); (iii) mimicked by 1,2-dioctanoylglycerol (DiC8) but not by 4 alpha-phorbol 12,13-didecanoate (which does not activate protein kinase C); (iv) attenuated by Staurosporine. These results indicate that activation of protein kinase C in platelets may provide a regulatory mechanism to abrogate the effects of the endogenous adenylate cyclase stimulant PGD2 without compromising the effects of exogenous stimulants of adenylate cyclase (PGI2, 6-keto-PGE1, adenosine).     

Prostaglandin I2 synthesis and elevation of cyclic AMP levels in 3T3 fibroblasts

Arachidonic acid and prostaglandin H2 elevate the levels of adenosine 3':5'-monophosphate (cyclic AMP) in Balb/c 3T3 fibroblasts. This effect was inhibited by 15-hydroperoxy-5,8,11,13-eicosatetraenoic acid, an inhibitor of prostaglandin I2 synthase (Claesson, H.-E., Lindgren, J.A. and Hammarstr!om, S. (1977) FEBS Lett. 81, 415-418). After addition of arachidonic acid to 3T3 cultures, cellular cyclic AMP levels and growth medium concentrations of 6-ketoprostaglandin F1 alpha (degradation product of prostaglandin I2) were quantitatively determined. The stimulatory effect of exogenously-added prostaglandin I2 on cellular cyclic AMP levels was also determined. The results indicate that the endogenous production of prostaglandin I2 is sufficient to explain the stimulatory action of arachidonic acid on cyclic AMP formation in 3T3 fibroblasts.