J. R. S. Hoult

Pharmacological and biochemical actions of simple coumarins : Natural products with therapeutic potential

1. More than 300 coumarins have been identified from natural sources, especially green plants. The pharmacological and biochemical properties and therapeutic applications of simple coumarins depend upon the pattern of substitution. More complex related compounds based on the coumarin nucleus include the dicoumarol/warfarin anticoagulants, aflatoxins and the psoralens (photosensitizing agents). 2. Coumarin itself (1,2-benzopyrone) has long-established efficacy in slow-onset long-term reduction of lymphoedema in man, as confirmed in recent double-blind trials against elephantiasis and postmastectomy swelling of the arm. The mechanism of action is uncertain, but may involve macrophage-induced proteolysis of oedema protein. However, coumarin has low absolute bioavailability in man (< 5%), due to extensive first-pass hepatic conversion to 7-hydroxycoumarin followed by glucuronidation. It may, therefore, be a prodrug. 3. Scoparone (6,7-dimethoxycoumarin) has been purified from the hypolipidaemic Chinese herb Artemisia scoparia and shown to reduce the proliferative responses of human peripheral mononuclear cells, to relax smooth muscle, to reduce total cholesterol and triglycerides and to retard the characteristic pathomorphological changes in hypercholesterolaemic diabetic rabbits. Various properties of scoparone were suggested to account for these findings, including ability to scavenge reactive oxygen species, inhibition of tyrosine kinases and potentiation of prostaglandin generation. 4. Osthole (7-methoxy-8-[3-methylpent-2-enyl]coumarin) from Angelica pubescens, used also in Chinese medicine, causes hypotension in vivo, and inhibits platelet aggregation and smooth muscle contraction in vitro. It may interfere with calcium influx and with cyclic nucleotide phosphodiesterases. 5. Cloricromene, a synthetic coumarin derivative, also possesses antithrombotic antiplatelet actions, inhibits PMN neutrophil function and causes vasodilatation. Some of these properties of cloricromene have been ascribed to inhibition of arachidonate release from membrane phospholipids. 6. Simple coumarins possessing ortho-dihydroxy functions, such as fraxetin and 4-methyldaphnetin, are potent inhibitors (low micromolar) of lipid peroxidation and scavengers of superoxide anion radicals and of aqueous alkylperoxyl radicals, but may be pro-oxidant (enhancing generation of hydroxyl radicals) in the presence of free iron ions. These coumarins also inhibit the proinflammatory 5-lipoxygenase enzyme at micromolar concentrations. Another related coumarin, 5,7-dihydroxy-4-methylcoumarin, is of special interest as it inhibits lipid peroxidation, and scavenges alkylperoxyl and superoxide radicals. Unlike most other simple coumarins studied, 5,7-dihydroxy-4-methylcoumarin also scavenges hypochlorous acid, and is a potent inhibitor of cyclo-oxygenase, but is not pro-oxidant. 7. 5,7- and 6,7-dihydroxy-4-methylcoumarin both reduced the duration of ventricular fibrillation in postischaemic reperfused isolated perfused rat hearts (in which oxygen-derived free radicals are implicated), showing that these antioxidant coumarins possess beneficial properties in this pathophysiological model. 8. In view of the established low toxicity, relative cheapness, presence in the diet and occurrence in various herbal remedies of coumarins, it appears prudent to evaluate their properties and applications further.

ACTIONS OF FLAVONOIDS AND COUMARINS ON LIPOXYGENASE AND CYCLOOXYGENASE

Publisher Summary This chapter discusses the actions of flavonoids and coumarins on lipoxygenase and cyclooxygenase. Flavonoids and coumarins are polyphenolic compounds containing multiple substituents on the benzo-γ-pyrone nucleus and are widely distributed in the plant kingdom. Reviews have emphasized that flavonoids and coumarins possess numerous biological and pharmacological properties, some of potential therapeutic interest. Various flavonoids and coumarins which possess anti-inflammatory actions in bioassays are identified in extracts prepared from plants used for this purpose in traditional folk medicines. The chapter describes methods used for the analysis of the actions of flavonoids and coumarins on the generation of eicosanoids by stimulated leukocytes. These cells are well suited for this purpose as they readily generate eicosanoids after engulfing opsonized particles or as a result of activation by soluble stimulants such as calcium ionophores, chemotactic peptides, and arachidonic acid itself. They are directly involved in the acute inflammatory process. Both rat mixed peritoneal leukocytes and human polymorphonuclear neutrophil leukocytes (PMNs) are used, although both cell types yield essentially similar results. The eicosanoid products of arachidonate metabolism are measured by radioimmunoassay or by radiometric (radio-TLC) methods, both of which are described in the chapter.

Repression of inducible nitric oxide synthase and cyclooxygenase-2 by prostaglandin E2 and other cyclic AMP stimulants in J774 macrophages

The enhanced nitric oxide (NO) and prostaglandin (PG) generation of activated macrophages is controlled by glucocorticoid-sensitive inducible nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2), respectively. Negative feedback regulation of iNOS expression by the products of both pathways has been suggested, but their effects on COX-2 expression have not been examined. We hae investigated the effect of E- and l-series prostaglandins that activate adenylate cyclase (AC), forskolin (a direct activator of AC), and other agents that influence the cyclicAMP/cyclicGMP systems on the ability of E. coli endotoxin (lipopolysaccharide, LPS) to induce iNOS and COX-2 in the murine macrophage cell line J774. After a 2-hr pretreatment before adding endotoxin, PGE2, PGI2, forskolin, IBMX (isobutylmethylxanthine, a cyclicAMP/cyclicGMP phosphodiesterase inhibitor), 8-bromo cyclicAMP, and arachidonic acid itself all inhibited the expression of both iNOS and COX-2 (as shown by Western blotting) and reduced NO release and COX activity, whereas PGF2 alpha and 8-bromo cyclic GMP were only weakly effective. The effects of PGE2, PGI2, and forskolin were enhanced by cotreatment with IBMX. The suppression of LPS-induced iNOS induction by PGE2 was functionally significant, in that it protected against the mild cytotoxicity of the NO generated in response to endotoxin. These results provide the first direct evidence for the feedback regulatory suppression of COX-2 induction by a PG-driven cAMP-mediated process, and show that the modulation of iNOS and COX-2 induction shares common features. They also suggest that such modulation is normally held in check by high phosphodiesterase activity within these cells.

Inhibition of 5-lipoxygenase and cyclo-oxygenase in leukocytes by feverfew. Involvement of sesquiterpene lactones and other components

Leaves or infusions of feverfew, Tanacetum parthenium, have long been used as a folk remedy for fever, arthritis and migraine, and derived products are widely available in U.K. health food shops. Previous reports have suggested interactions with arachidonate metabolism. Crude chloroform extracts of fresh feverfew leaves (rich in sesquiterpene lactones) and of commercially available powdered leaves (lactone-free) produced dose-dependent inhibition of the generation of thromboxane B2 (TXB2) and leukotriene B4 (LTB4) by ionophore- and chemoattractant-stimulated rat peritoneal leukocytes and human polymorphonuclear leukocytes. Approximate IC50 values were in the range 5-50 micrograms/mL, and inhibition of TXB2 and LTB4 occurred in parallel. Isolated lactones (parthenolide, epoxyartemorin) treated with cysteine (to neutralize reactive alpha-methylene butyrolactone functions of the sesquiterpenes). Inhibition of eicosanoid generation appeared to be irreversible but not time-dependent. We conclude that feverfew contains a complex mixture of sesquiterpene lactone and non-sesquiterpene lactone inhibitors of eicosanoid synthesis of high potency, and that these biochemical actions may be relevant to the claimed therapeutic actions of the herb.

Superoxide scavenging activity in leukocytes and absence of cellular toxicity of a series of coumarins.

Sixteen synthetic or plant-derived coumarins of dietary importance with different patterns of substitution were tested for their capacity to scavenge superoxide and for their cytotoxicity. Superoxide was generated by human polymorphonuclear leukocytes stimulated by phorbol myristate acetate and was measured using the reduction of ferricytochrome c or of nitrobule tetrazolium (NBT). Eleven of the coumarins, all lacking dihydroxy substitution, did not scavenge superoxide. Of the remaining five, the most potent scavenger was fraxetin (7,8-dihydroxy-6-methoxycoumarin) with an IC50 (concentration producing 50% inhibition) of 2.3 microM in the cytochrome assay and 5.8 microM using NBT. The other four coumarins (all containing ortho-dihydroxy catechol functions, and found previously to be pro-oxidant in cell-free systems by virtue of reduction of ferric to ferrous ions), themselves rapidly reduced cytochrome c. Therefore their effects on superoxide were measured using NBT, yielding IC50 values in the range 8.5 to 82.0 microM. Fraxetin and the other active and inactive coumarins were not directly cytotoxic at 100 microM to leukocytes or to erythrocytes, as shown by their failure to cause release of cytosolic lactate dehydrogenase or to cause haemolysis, respectively. However, all five dihydroxylated pro-oxidant coumarins were toxic to NS20Y neuroblastoma cells in 24 hr culture, whereas the other eleven coumarins were nontoxic. We conclude that 7,8-dihydroxylated coumarins such as fraxetin are agents which are not themselves directly cytotoxic and are capable of direct scavenging of superoxide anion radicals, an action which might be protective at sites of leukocyte activation during inflammation. However, in the presence of free ferric ions they may exert potentially damaging pro-oxidant actions, including cytotoxicity. This series of compounds provides a useful basis for structure-activity studies designed to achieve separation or combination of these properties.

Cytotoxicity to macrophages of tetrandrine, an antisilicosis alkaloid, accompanied by an overproduction of prostaglandins

Tetrandrine, an anti-inflammatory immunosuppressive bisbenzylisoquinoline alkaloid of Chinese herbal origin, is widely used to treat silicosis and interferes with the regulation of calcium in many cell types. We investigated its effect on the cellular integrity of macrophages and on their ability to generate prostaglandins and nitric oxide, mediators of inflammation with immunomodulatory roles. Tetrandrine at 10(-7) M to 10(-4) M caused dose- and time-dependent loss of cell viability of mouse peritoneal macrophages, guinea-pig alveolar macrophages and mouse macrophage-like J774 cells. Loss of viability (50%) occurred within 1-3 hr and required approximately 5 x 10(-6) M tetrandrine. Loss of macrophage viability after tetrandrine treatment was accompanied by the generation of large amounts of prostaglandin E2 (PGE2), to levels 285-877% of control. Coincubation with indomethacin abolished PGE2 generation, but did not prevent cell death. Tetrandrine did not cause generation of nitric oxide. Verapamil also reduced the viability of mouse peritoneal macrophages and J774 cells, but did not cause PGE2 overproduction, except at 10(-4) M in mouse peritoneal macrophages. In macrophages cultured with lipopolysaccharide and interferon-gamma to induce the generation of large amounts of both PGE2 and nitric oxide, tetrandrine reduced mediator release and their forming enzymes (cyclo-oxygenase-2 and inducible nitric oxide synthase), secondary to cytotoxicity. The predominant action of tetrandrine is to exert a cytotoxic effect on macrophages, perhaps by interfering with calcium homeostasis; this leads to overproduction of immunomodulatory but proinflammatory prostaglandin. This may be relevant to its protective actions in human fibrosing silicosis, in which there is alveolar macrophage involvement.

Groups I, II and III extracellular phospholipases A2: selective inhibition of group II enzymes by indomethacin but not other NSAIDs.

The three types (groups I, II and III) of stable extracellular 14 kDa phospholipase A2 enzymes differ in their primary amino acid sequences and their properties. It may thus be possible to design low-molecular weight inhibitors targeted to the secretory form of mammalian PLA2. this enzyme has been implicated in inflammatory disorders. We have studied the inhibition of four distinct PLA2 enzymes by a range of NSAIDs, using3H-oleate release from prelabelled membranes ofE. coli for assay. The enzymes used were cobra venom PLA2 (Naja naja, a group I enzyme), bee venom PLA2 (Apis mellifera, group III), recombinant human synovial PLA2 (group II) and rat peritoneal PLA2 (group II). Under the conditions of the3H-oleateE. coli assay, 1 mM concentrations of aspirin, sodium salicylate, paracetamol (acetaminophen), oxphenbutazone, ibuprofen, flurbiprofen and nabumetone failed to inhibit significantly any of the four enzymes. However, indomethacin inhibited all four enzymes, although effects were greatest on the two group II enzymes (rat peritoneal and human synovial PLA2). Approximate IC50 values were 28 and 35 μM, respectively. Inhibition by indomethacin was not time dependent and was greater at micromolar rather than millimolar levels of calcium. We conclude that indomethacin but not the other tested classes of NSAID inhibits the group II PLA2 enzyme in a selective manner and suggest that this may be relevant both to its clinical spectrum and to the design of novel pharmaceutical leads.

Calcium overload toxicity and functional impairment in peritoneal leukocytes elicited by glycogen or interleukin-1β

Although calcium plays an important role in the activation of leukocytes for such processes as eicosanoid biosynthesis, secretion of granular constituents and superoxide generation, sustained high levels of intracellular calcium ions may be toxic. We have previously found that high concentrations of calcium ionophores induce a rapid-onset “calcium overload” toxicity in rat peritoneal leukocytes, in which functional responses such as β-glucuronidase secretion, superoxide generation and leukotriene B4 synthesis are greatly attenuated, and some leakage of cytoplasmic LDH occurs. We have now compared this phenomenon in peritoneal leukocytes elicited from animals pretreated in three ways: glycogen, interleukin-1β (IL-1β) alone or glycogen plus IL-1β. Peritoneal administration of IL-1β caused elicitation of cells which were enriched in eosinophils; however, the functional responses of the cells in all three groups were broadly similar in terms of the ability of the agonists FMLP, PMA and A23187 to initiate superoxide generation, β-glucuronidase secretion and leukotriene generation. Cells from all three treatment groups showed diminished responsiveness at 10−5M A23187, indicative of calcium overload toxicity. This was most evident for the superoxide and β-glucuronidase responses. Some quantitative differences observed between treatment groups may reflect the different sensitivities of the various cells contained in the mixed leukocyte preparations. We conclude that IL-1β induces leukocyte emigration into the peritoneal cavity but that the cell population is different from that induced by glycogen. However, the cells retain susceptibility to calcium overload toxicity.

Failure of anti-inflammatory steroids to inhibit prostaglandin release from the hydronephrotic rabbit kidney

The release of prostaglandins E2, F2α, I2 and thromboxane A2 from isolated perfused normal and hydronephrotic rabbit kidneys was investigated by extraction and radio-immunoassay. In both types of kidneys, basal PG efflux increased with time and was not altered by co-perfusion with dexamethasone or hydrocortisone. Several vasoactive substances at 1 to 4 μg (e.g., bradykinin, angiotensin II, substance P, noradrenaline and vasopressin) caused release of additional amounts of prostaglandins. PGE2 and 6-keto PGF1α were the major prostanoids detected, but substantial amounts of PGF2α were also found. Thromboxane A2 was not released from normal kidneys. In hydronephrotic kidneys there was greatly augmented release of prostaglandins E2 and I2, some increases in PGF2α, and the appearance of substantial amounts of thromboxane A2 (measured as immunoreactive TXB2) when the kidneys were challenged with angiotensin, bradykinin and vasopressin, and smaller augmentation of the response to noradrenaline and substance P. There was no evidence that these evoked increases in renal PG output could be inhibited by dexamethasone or hydrocortisone. Some explanations for the failure of steroids to alter prostanoid metabolism from arachidonate in rabbit kidney are discussed, and it is proposed that there are clear exceptions to the concept that steroids inhibit prostaglandin generation in intact tissues.