Discovery and biological characterization of a novel series of androgen receptor modulators.

BACKGROUND AND PURPOSE: Selective androgen receptor modulators are of great value in the treatment of prostate cancer. The purpose of this study was to provide a preliminary characterization of a new class of non-steroidal androgen receptor modulators discovered in a high-throughput screening campaign. EXPERIMENTAL APPROACH: Competitive receptor binding, luciferase-based reporter methods, cell proliferation and in vivo assays were employed to evaluate an initial set of compounds from chemistry efforts. KEY RESULTS: Forty-nine analogues from the chemistry efforts showed high affinity binding to androgen receptors, agonist and/or antagonist activities in both CV-1 and MDA-MB-453 transfection assays. A proliferation assay in LNCaP cells also exhibited this profile. A representative of these non-steroidal compounds (compound 21) was devoid of activity at other nuclear receptors (oestrogen, progesterone, glucocorticoid and mineralocorticoid receptors) in the CV-1 co-transfection assay. At the same time, in an immature castrated rat model, it behaved as an androgen receptor antagonist against the growth of prostate, seminal vesicles and levator ani induced by exogenous androgen. Separation of compound 21 into its enantiomers showed that nearly all the androgen receptor modulating activity and binding resided in the dextrorotatory compound (23) while the laevorotatory isomer (22) possessed weak or little effect depending on the cell type studied. CONCLUSIONS AND IMPLICATIONS: These non-steroidal compounds may represent a new class of androgen receptor modulators for the treatment of not only prostate cancer but other clinical conditions where androgens and androgen receptors are involved in the pathological processes.

A selective peroxisome proliferator-activated receptor-gamma (PPARgamma) modulator blocks adipocyte differentiation but stimulates glucose uptake in 3T3-L1 adipocytes.

Peroxisome proliferator-activated receptor-gamma (PPARgamma) agonists such as the thiazolidinediones are insulin sensitizers used in the treatment of type 2 diabetes. These compounds induce adipogenesis in cell culture models and increase weight gain in rodents and humans. We have identified a novel PPARgamma ligand, LG100641, that does not activate PPARgamma but selectively and competitively blocks thiazolidinedione-induced PPARgamma activation and adipocyte conversion. It also antagonizes target gene activation as well as repression in agonist-treated 3T3-L1 adipocytes. This novel PPARgamma antagonist does not block adipocyte differentiation induced by a ligand for the retinoid X receptor (RXR), the heterodimeric partner for PPARgamma, or by a differentiation cocktail containing insulin, dexamethasone, and 1-methyl-3-isobutylxanthine. Surprisingly, LG100641, like the PPARgamma agonist rosiglitazone, increases glucose uptake in 3T3-L1 adipocytes. Such selective PPARgamma antagonists may help determine whether insulin sensitization by thiazolidinediones is mediated solely through PPARgamma activation, and whether there are PPARgamma-ligand-independent pathways for adipocyte differentiation. If selective PPARgamma modulators block adipogenesis in vivo, they may prevent obesity, lower insulin resistance, and delay the onset of type 2 diabetes.

Nonsteroidal progesterone receptor antagonists based on 6-thiophenehydroquinolines.

Synthesis and biological evaluation of 6-thiophene 1,2-dihydro or 1,2,3,4-tetrahydroquinoline derivatives resulted in a number of potent nonsteroidal antiprogestins.

Novel nonsecosteroidal vitamin D mimics exert VDR-modulating activities with less calcium mobilization than 1,25-dihydroxyvitamin D3.

BACKGROUND: The secosteroid 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) acts through the vitamin D receptor (VDR) to elicit many activities that make it a promising drug candidate for the treatment of a number of diseases, including cancer and psoriasis. Clinical use of 1,25(OH)2D3 has been limited by hypercalcemia elicited by pharmacologically effective doses. We hypothesized that structurally distinct, nonsecosteroidal mimics of 1,25(OH)2D3 might have different activity profiles from vitamin D analogs, and set out to discover such compounds by screening small-molecule libraries. RESULTS: A bis-phenyl derivative was found to activate VDR in a transactivation screening assay. Additional related compounds were synthesized that mimicked various activities of 1,25(OH)2D3, including growth inhibition of cancer cells and keratinocytes, as well as induction of leukemic cell differentiation. In contrast to 1, 25(OH)2D3, these synthetic compounds did not demonstrate appreciable binding to serum vitamin D binding protein, a property that is correlated with fewer calcium effects in vivo. Two mimics tested in mice showed greater induction of a VDR target gene with less elevation of serum calcium than 1,25(OH)2D3. CONCLUSIONS: These novel VDR modulators may have potential as therapeutics for cancer, leukemia and psoriasis with less calcium mobilization side effects than are associated with secosteroidal 1,25(OH)2D3 analogs.

5-Aryl-1,2,3,4-tetrahydrochromeno[3,4-f]quinolin-3-ones as a novel class of nonsteroidal progesterone receptor agonists: effect of A-ring modification.

Optimization of the 1,2-dihydroquinoline A-ring of a nonsteroidal human progesterone receptor (hPR) agonist pharmacophore (1) was performed by using the cotransfection and receptor binding assays as guides. The 3-keto group was discovered to regain the potent agonist activity which was lost upon removal of the 3,4-olefin, and it led to a novel hPR agonist series, 5-aryl-1,2,3,4-tetrahydrochromeno[3, 4-f]quinolin-3-ones. The new progestins demonstrated potent hPR agonist activity in the cotransfection assay and high binding affinity similar to progesterone. T47D human breast cancer cell line was employed for further characterization of the new progestins and a number of reference analogues. It was found that the new 3-keto analogues showed full agonist activity in the T47D assay, while the reference compounds from other related nonsteroidal hPR agonist series exhibited only partial agonist activity.

Estrogen receptor beta activates the human retinoic acid receptor alpha-1 promoter in response to tamoxifen and other estrogen receptor antagonists, but not in response to estrogen.

Human estrogen receptor-alpha (hERalpha) or -beta (hERbeta) transfected into Hep G2 or COS1 cells each responded to estrogen to increase transcription from an estrogen-responsive element (ERE)-driven reporter vector with similar fold induction through a classical mechanism involving direct receptor binding to DNA. ER antagonists inhibited this estrogen induction through both hERalpha and hERbeta, although raloxifene was more potent through ERalpha than ERbeta, and tamoxifen was more potent via ERbeta than ERalpha. We have shown previously that estrogen stimulated the human retinoic acid receptor-alpha-1 (hRARalpha-1) promoter through nonclassical EREs by a mechanism that was ERalpha dependent, but that did not involve direct receptor binding to DNA. We show here that in contrast to hERalpha, hERbeta did not induce reporter activity driven by the hRARalpha-1 promoter in the presence of estrogen. While hERbeta did not confer estrogen responsiveness on this promoter, it did elicit transcriptional activation in the presence of 4-hydroxytamoxifen (4-OH-Tam). Additionally, this 4-OH-Tam agonist activity via ERbeta was completely blocked by estrogen. Like ERalpha, transcriptional activation of this promoter by ERbeta was not mediated by direct receptor binding to DNA. While hERalpha was shown to act through two estrogen-responsive sequences within the promoter, hERbeta acted only at the 3'-region, through two Sp1 sites, in response to 4-OH-Tam. Other ER antagonists including raloxifene, ICI-164,384 and ICI-182,780 also acted as agonists through ERbeta via the hRARalpha-1 promoter. Through the use of mutant and chimeric receptors, it was shown that the 4-OH-Tam activity via ERbeta from the hRARalpha-1 promoter in Hep G2 cells required the amino-terminal region of ERbeta, a region that was not necessary for estrogen-induced ERbeta activity from an ERE in Hep G2 cells. Additionally, the progesterone receptor (PR) antagonist RU486 acted as a weak (IC50 >1 microM) antagonist via hERalpha and as a fairly potent (IC50 approximately 200 nM) antagonist via hERbeta from an ERE-driven reporter in cells that do not express PR. Although RU486 bound only weakly to ERalpha or ERbeta in vitro, it did bind to ERbeta in whole-cell binding assays, and therefore, it is likely metabolized to an ERbeta-interacting compound in the cell. Interestingly, RU486 acted as an agonist through ERbeta to stimulate the hRARalpha-1 promoter in Hep G2 cells. These findings may have ramifications in breast cancer treatment regimens utilizing tamoxifen or other ER antagonists and may explain some of the known estrogenic or antiestrogenic biological actions of RU486.

Expression, characterization, and purification of C-terminally hexahistidine-tagged thromboxane A2 receptors.

Thromboxane A2 (TxA2) receptors belong to the class of G-protein-coupled receptors. Knowledge of the relationship of structure to function for TxA2 receptors is limited because of their low levels of expression, lengthy purification procedures and poor recoveries. A C-terminal hexahistidine-tag (C-His) was ligated to the alpha-isoform of TxA2 receptors and expressed in COS-7 and Chinese hamster ovary cells. The C-His-TxA2 receptors bound the radioligands 125I-7-[(1R,2S,3S,5R)-6, 6-dimethyl-3-(4-benzenesulfonylamino)bicyclo[3.1. 1]hept-2-yl]-5(Z)-heptenoic acid, an antagonist, and 125I-[1S-1alpha, 2beta(5Z),3alpha(1E,3S*), 4alpha]-7-[3[(3-hydroxy-4-(4'-phenoxy)-1butenyl)-7-oxabicycl o-[2.2. 1]heptan-2-yl]-5-heptanoic acid, an agonist, with affinities not significantly different from those of the wild type (wt)-TxA2 receptors. LipofectAMINE transfection of the cDNAs resulted in high levels of expression (Bmax = 95 +/- 6 pmol/mg) of the C-His-TxA2 receptors. In competition binding studies the IC50 values of five different ligands were not significantly different between C-His-TxA2 and wt-TxA2 receptors. Agonist-induced stimulation of cAMP and total inositol phosphate formation were not significantly different between the two receptors. Purification on a Ni2+-NTA column resulted in a rapid (within 4 h) purification with a 36 +/- 2% recovery and a 30 +/- 6-fold purification (n = 5). The partially purified receptors were resolved on SDS-polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, dissolved in acetone/trifluoroacetic acid/hexafluoroisopropanol/sinapinic acid, and successfully subjected to matrix-assisted laser desorption ionization-time of flight mass spectrometry analysis. The results suggest that the combination of a high level of expression of C-His-TxA2 receptors and a rapid purification procedure followed by SDS- polyacrylamide gel electrophoresis may provide a useful approach for mass-spectrometry based structure-function and other studies of TxA2 receptors.

5-Aryl-1,2-dihydrochromeno[3,4-f]quinolines: a novel class of nonsteroidal human progesterone receptor agonists.

The development of a novel class of nonsteroidal human progesterone receptor (hPR) agonists, 5-aryl-1,2-dihydro-5H-chromeno[3,4-f]quinolines 2, is described. The introduction of a 5-aryl group into the 1,2-dihydrocoumarino[3,4-f]quinoline core 1 is the key for progestational activities. The structure-activity relationship (SAR) studies of the 5-aryl substituents generated a series of potent hPR agonists, which exhibited similar biological activity (EC50 = 8-30 nM) to the natural hormone progesterone (EC50 = 2.9 nM) in cell-based assays with efficacies ranging from 28% to 96%. Most of the analogues displayed similar or greater binding affinity (Ki = 0.41-3.6 nM) than progesterone (Ki = 3.5 nM). Three representative analogues (13, 15, and 24) demonstrated in vivo activities in mammary gland morphology/uterine wet weight assay in ovariectomized rats.

5-Aryl-1,2-dihydro-5H-chromeno[3,4-f]quinolines as potent, orally active, nonsteroidal progesterone receptor agonists: the effect of D-ring substituents.

Several 5-(4-chlorophenyl)-1,2-dihydro-5H-chromeno[3,4-f]quinolines were prepared to determine the effects of substitution at C(8) and C(9) on the progestational activity of this pharmacophore. In combination with a halogen (F or Cl) at C(9), replacement of the C(5) aryl group with variously substituted aryl groups resulted in optimization of the progestational activity, affording compounds with in vitro activity greater than that of progesterone as measured by a cotransfection assay using human progesterone receptor subtype-B (hPR-B). Binding affinities (Ki) to hPR-A were subnanomolar in many cases. These in vitro effects were verified in vivo using a rodent model. Compound 10 (LG120794, 9-chloro-5-(4-chlorophenyl)-1,2-dihydro-2,2,4-trimethyl-5H-chromeno++ +[3,4-f] quinoline) was more potent than medroxyprogesterone acetate at counterpoising the effects of estradiol benzoate in the uterine wet weight assay using immature rats.

Discovery of novel retinoic acid receptor agonists having potent antiproliferative activity in cervical cancer cells.

Retinoic acid receptor (RAR) active retinoids have proven therapeutically useful for treating certain cancers and dermatological diseases. Herein, we describe the discovery of two new RAR active trienoic acid retinoids, (2E,4E,6E)-7-(3,5-di-tert-butylphenyl)-3-methylocta-2, 4,6-trienoic acid (10a, ALRT1550) and (2E,4E,6Z)-7-(3,5-di-tert-butylphenyl)-3-methylocta-2, 4,6-trienoic acid (10b, LG100567). ALRT1550 is a RAR selective retinoid which exhibits exceptional potency in both competitive binding and cotransfection assays. Moreover, it is the most potent antiproliferative retinoid described to date and thus has implications for the treatment of certain cancers. LG100567 is a potent panagonist which activates both RARs and retinoid X receptors.

Synthesis and biological activity of novel nonsteroidal progesterone receptor antagonists based on cyclocymopol monomethyl ether.

A novel class of nonsteroidal progesterone receptor antagonists has been synthesized and was shown to exhibit moderate binding affinity for hPR-A, the ability to inhibit the transcriptional activity of human progesterone receptor (hPR) in cell-based assays, and anti-progestational activity in a murine model. Cyclocymopol monomethyl ether, a component of the marine alga Cymopolia barbata was weakly active in random screening against PR. Investigations into the SAR surrounding the core of this natural product lead structure resulted in improved in vitro activity. In contrast to the cross-reactivity profiles observed with known steroidal antiprogestins, compounds of the general structural class described display a high degree of selectivity for the progesterone receptor and no functional activity on the glucocorticoid receptor.

Discovery of novel intracellular receptor modulating drugs.

Utilizing the co-transfection assay as a guide to determining structure activity relationships, we have been pursuing the discovery of non-steroidal hPR modulators. Small molecule, non-steroidal lead structures have been identified. Optimization of these structures has yielded more potent hPR modulators. Improved cross-reactivity profiles with other intracellular receptors are a feature of these compounds owing to their non-steroidal nature.

Design and synthesis of potent retinoid X receptor selective ligands that induce apoptosis in leukemia cells.

Structural modifications of the retinoid X receptor (RXR) selective compound 4-[1-(3,5,5,8,8-pentamethyl-5,6,7,8-tetrahydro-2- naphthyl)ethenyl]benzoic acid (LGD1069), which is currently in phase I/IIA clinical trials for cancer and dermatological indications, have resulted in the identification of increasingly potent retinoids with > 1000-fold selectivity for the RXRs. This paper describes the design and preparation of a series of RXR selective retinoids as well as the biological data obtained from cotransfection and competitive binding assays which were used to evaluate their potency and selectivity. The most potent and selective of the analogs is 6-[1-(3,5,5,8,8-pentamethyl-5,6,7,8-tetrahydronaphthalen-2- yl)cyclopropyl]nicotinic acid (12d; LG100268). This compound has proven useful for investigating RXR dependent biological pathways including the induction of programmed cell death (PCD) and transglutaminase (TGase) activity. Our studies indicate that the induction of PCD and TGase in human leukemic myeloid cells is dependent upon activation of RXR-mediated pathways.

Specific interactions of progestins and anti-progestins with progesterone antibodies, plasma binding proteins and the human recombinant receptor.

This structure-activity study compares the affinity of a series of progestins, progesterone metabolites and anti-progestins for a panel of monoclonal antibodies to progesterone, coypu (Myocastor coypus) or guinea pig plasma progesterone-binding proteins (PPBPs) and the human recombinant progesterone receptor A form (PR-A). The compounds tested were progesterone, Promegestone (R5020), Mifepristone (RU486), ZK98,734, Onapristone (ZK98,299), 11 alpha-hydroxyprogesterone, 11 alpha-progesterone hemisuccinate, androsterone, etiocholanolone, 5 alpha- and 5 beta-pregnane-3,20-diones, and 20 alpha- and 20 beta-hydroxyprogesterones. The Ki values for these ligands were determined by competitive binding assays using radiolabelled progesterone as the binding site ligand. For anti-progesterone antibodies (e.g. DB3 and 11/32), only progesterone (3.6-8.8 nM), the 11 alpha-derivatives (1.0-5.5 nM) used to prepare the immunogen and the two 5-pregnanediones (20.9-45.1 nM) were bound with high affinity. For PR-A, high affinity binding was found with receptor agonists (Ki = 1.1-6.2 nM), both 5- and 20-reduced metabolites, and antagonists (0.6-28.0 nM), but not with the 11 alpha-derivatives (950 nM-1.0 microM). In contrast, the PPBPs displayed high affinity interactions with progesterone (3.5-4.2 nM) and both 5 alpha- and 20 alpha-reduced metabolites (2.4-3.4 nM). Binding with the beta-isomers and R5020 was less pronounced (22-170 nM) and there was no evidence of high affinity binding with PR antagonists (> 1.0 microM). Analogs with the 17-keto group did not bind to any of the binders studied. Thus, commonalities among the three types of protein binders were their comparable binding affinities for progesterone (3.5-8.8 nM) and 5-pregnanedione isomers (2.4-330 nM), and a lack of binding for two C17-keto steroids (androsterone and etiocholanolone). The results imply that the tertiary features of the binding domain of these three types of proteins are sufficiently different to result in unique binding structures.

Anti-progesterone antibody administration and the impairment of postpartum maternal care in mice.

Passive transfer of a monoclonal antibody against progesterone produces a high incidence of maternal rejection in mice after recovery from antibody-induced infertility. To investigate the mechanisms involved in this reduction of maternal care, we have examined whether the effect is due to long-term exposure to antibody. Antibody was administered i.p. either on day 2 or day 17 of pregnancy. When a low dose (1.0 nmol) was given on day 2, pregnancy proceeded normally but 44.8% pups delivered at term were rejected compared with 12.7% in the control group. When a higher dose (4.5 nmol) of antibody was given on day 17, pregnancy continued normally to term and the rejection rate was 48.8% (control: 11.1%). When the same amount of antibody was injected after delivery (day 1 of lactation), no detrimental effect was found on subsequent maternal care to the young, the rejection rate being comparable between antibody-treated and control groups (5.3% vs 4.6%). To determine if the presence of antibody interfered with lactation or suckling, a bolus injection of 10 microCi [3H]H2O was given to mice treated at day 17 with antibody or saline. The levels of radioactivity present in both mothers and pups and the first 5-day pup growth curves showed identical patterns, indicating that milk availability and the suckling process were not affected. Crossfostering studies revealed that antibody-treated mothers rejected 25.5% of fostered pups compared with 8.5% found in the control females when antibody was administered on day 17 of pregnancy and the entire litters were crossfostered between the two groups immediately after delivery.(ABSTRACT TRUNCATED AT 250 WORDS)

Nonsteroidal human progesterone receptor modulators from the marine alga Cymopolia barbata.

The co-transfection assay is a novel functional assay using cells transiently transfected with plasmids encoding intracellular receptors and corresponding reporter genes. Using this assay, natural product extracts were tested to identify compounds that modulate intracellular receptor activity, measured as changes in reporter gene activity. A crude extract of the marine alga Cymopolia barbata was found to inhibit progesterone-stimulated reporter gene expression in cells transfected with the human progesterone receptor (hPR) and an appropriate reporter construct. Purification of the active constituents of the extract, guided by the co-transfection assay, yielded two diastereomers of cyclocymopol monomethyl ether, possessing opposing pharmacological activities with the hPR. The antagonist (3R)-cyclocymopol monomethyl ether (LG100127) blocked 1 nM progesterone-stimulated reporter gene expression with an IC50 value of 549 +/- 55 nM in the co-transfection assay. The agonist (3S)-cyclocymopol monomethyl ether (LG100128) had efficacy similar to that of progesterone and an EC50 value of 35 +/- 2 nM. Stimulation by progesterone of the hPR in the human breast cancer cell line T-47D results in enhanced expression of alkaline phosphatase; LG100127 blocked alkaline phosphatase expression stimulated either by progesterone or by LG100128, and LG100128 mimicked progesterone in this assay. Both diastereomers displaced [3H]progesterone from baculovirus-expressed hPR. LG100127 and LG100128 each interacted with the human androgen receptor but did not interact with the human glucocorticoid receptor, estrogen receptor, vitamin D receptor, or retinoid receptors. In summary, these in vitro studies describe the first nonsteroidal pharmacophores for the hPR and demonstrate the use of the co-transfection assay in their discovery.

The A and B isoforms of the human progesterone receptor operate through distinct signaling pathways within target cells.

The biological response to progesterone is mediated by two distinct forms of the human progesterone receptor (hPR-A and hPR-B). In most cell contexts, hPR-B functions as a transcriptional activator of progesterone-responsive genes, whereas hPR-A functions as a transcriptional inhibitor of all steroid hormone receptors. We have created mutations within the carboxyl terminus of hPR which differentially effect the transcriptional activity of hPR-B in a cell- and promoter-specific manner. Analogous mutations, when introduced into hPR-A, have no effect on its ability to inhibit the transcriptional activity of other steroid hormone receptors. The observed differences in the structural requirements for hPR-B and hPR-A function suggest that transcriptional activation and repression by PR are mediated by two separate pathways within the cell. In support of this hypothesis, we have shown that hPR-A mediated repression of human estrogen receptor (hER) transcriptional activity is not dependent on hER expression level but depends largely on the absolute expression level of hPR-A. Thus, it appears that hPR-A inhibits hER transcriptional activity as a consequence of a noncompetitive interaction of hPR-A with either distinct cellular targets or different contact sites on the same target. We propose that hPR-A expression facilitates a ligand-dependent cross-talk among sex steroid receptor signaling pathways within the cell. It is likely, therefore, that alterations in the expression level of hPR-A or its cellular target can have profound effects on the physiological or pharmacological responses to sex steroid hormone receptor ligands.

Synthesis and structure-activity relationships of novel retinoid X receptor-selective retinoids.

Two series of potent retinoid X receptor (RXR)-selective compounds were designed and synthesized based upon recent observation that (E)-4-[2-(5,5,8,8-tetramethyl-5,6,7,8-tetrahydro-2-naphthalenyl)-1- propenyl]benzoic acid (TTNBP) binds and transactivates only the retinoic acid receptor (RAR) subtypes whereas (E)-4-[2-(3,5,5,8,8-pentamethyl-5,6,7,8- tetrahydro-2-naphthalenyl)-1-propenyl]benzoic acid (3-methyl TTNPB) binds and transactivates both the RAR and RXR subfamilies. Addition of functional groups such as methyl, chloro, bromo, or ethyl to the 3 position of the tetrahydronaphthalene moiety of 4-[(5,5,8,8-tetramethyl-5,6,7,8-tetrahydro-2-naphthyl)carbonyl]benzoic acid (5a) and 4-[1-(5,5,8,8-tetramethyl-5,6,7,8-tetrahydro-2- naphthyl)ethenyl]benzoic acid (6a) results in compounds which elicit potent and selective activation of the RXR class. Such RXR-selective compounds offer pharmacological tools for elucidating the biological role of the individual retinoid receptors with which they interact. Activation profiles in cotransfection and competitive binding assays as well as molecular modeling calculations demonstrate critical structural determinants that confer selectivity for members of the RXR subfamily. The most potent compound of these series, 4-[1-(3,5,5,8,8-pentamethyl-5,6,7,8-tetrahydro-2-naphthyl)ethenyl]ben zoi c acid (6b), is the first RXR-selective retinoid (designated as LGD1069) to enter clinical trials for cancer indications.

Biochemical and pharmacological activity of arene-fused prostacyclin analogues on human platelets.

Human platelets have been employed as an assay system to evaluate the pharmacological activity of a group of stable, arene-fused prostacyclin analogs. Prostacyclin (PGI2) is a highly active member of the eicosanoid family and is relatively unstable under physiological conditions. Prostacyclin's best characterized activities are those of inhibition of platelet aggregation and relaxation of vascular smooth muscle. These activities are mediated in large part via elevation of intracellular levels of cyclic AMP subsequent to receptor occupation and activation of adenylate cyclase. We previously described the synthesis of a series of arene-fused prostacyclin analogs with stability in aqueous media at physiological pH. Several of these compounds have prostacyclin-like activities, i.e., competitive binding at the platelet prostacyclin receptor, elevation of intraplatelet cyclic AMP levels and inhibition of human platelet aggregation. One compound in particular (11a) demonstrated these activities with potency similar to PGI2, i.e., Kd at platelet receptor of 3.7 nM and IC50 for inhibition of collagen-induced human platelet aggregation in plasma of 2.9 nM.