Chapter 33: Lipid-Derived Autacoids: Eicosanoids and Platelet-Activating Factor

History. In 1930, Kurzrok and Lieb, two American gynecologists, observed that strips of uterine myometrium relax or contract when exposed to semen. Subsequently, Goldblatt in England and von Euler in Sweden reported independently on smooth muscle-contracting and vasodepressor activities in seminal fluid and accessory reproductive glands. In 1935, von Euler identified the active material as a lipid-soluble acid, which he named prostaglandin, inferring its origin in the prostatic gland. Samuelsson, Bergström, and their colleagues elucidated the structures of prostaglandin E₁ (PGE₁) and prostaglandin F₁α (PGF₁α) in 1962. In 1964, Bergström and coworkers, and van Dorp and associates, independently achieved biosynthesis of PGE₂ from arachidonic acid (AA). Discovery of TxA₂, PGI₂, and the LTs followed. Vane, Smith, and Willis reported that aspirin and NSAIDs act by inhibiting prostaglandin biosynthesis (Vane, 1971). This remarkable period of discovery linked the Nobel Prize of von Euler in 1970 to that of Bergström, Samuelsson, and Vane in 1982.

PGs, LTs, and related compounds are called eicosanoids, from the Greek eikosi ("twenty"). Precursor essential fatty acids contain 20 carbons and three, four, or five double bonds: 8,11,14-eicosatrienoic acid (dihomo-γ-linolenic acid), 5,8,11,14-eicosatetraenoic acid [AA; Figure 33–1], and 5,8,11,14,17-eicosapentaenoic acid (EPA). In humans, AA, the most abundant precursor, is either derived from dietary linoleic acid (9,12-octadecadienoic acid) or ingested directly as a dietary constituent. EPA is a major constituent of oils from fatty fish such as salmon.

Products of Prostaglandin G/H Synthases. Synthesis of the PGs, PGI₂, and thromboxane, collectively termed prostanoids, is accomplished in a stepwise manner by a complex of microsomal enzymes. Successive metabolism of AA by the COX and hydroperoxidase (HOX) activities of the PG endoperoxide G/H synthases, generate the cyclic endoperoxide PGs G and H (Figure 33–1). Isomerases and synthases affect the transformation of prostaglandin H₂ (PGH₂) into terminal prostanoids distinguished by substitutions on their cyclopentane rings.

PGs of the E and D series are hydroxyketones, whereas the F_α PGs are 1,3-diols (Figure 33–1). A, B, and C PGs are unsaturated ketones that arise nonenzymatically from PGE during extraction procedures; it is unlikely that they occur biologically. Prostaglandin J₂ (PGJ₂) and related compounds result from the dehydration of prostaglandin D₂ (PGD₂). Prostaglandin I₂ (PGI₂, prostacyclin) has a double-ring structure; in addition to a cyclopentane ring, a second ring is formed by an oxygen bridge between carbons 6 and 9. Thromboxanes (TXs) contain a six-member oxirane ring, instead of the cyclopentane ring of the PGs. Levuglandins are formed by non-enzymatic rearrangement of PGH₂. These γ-ketoaldehydes are highly reactive so that they are detected in vivo as their protein adducts (Salomon et al., 1997). The main classes are further subdivided in accord with the number of double bonds in their side chains, as indicated by numerical subscripts. Dihomo-γ-linolenic acid is the precursor of the one series, AA for the two series, and EPA for the three series. Prostanoids derived from AA carry the subscript 2 and are the major series in mammals. There is little evidence that one- or three-series prostanoids are made in adequate amounts to be important under normal circumstances. However, the health benefits of dietary supplementation with ω-3 fatty acids, like EPA or the 22-carbon-containing docosahexaenoic acid (DHA), remain a focus of investigation.

The first enzyme in the prostanoid synthetic pathway is PG endoperoxide G/H synthase, which is colloquially called cyclooxygenase, or COX. There are two distinct COX isoforms, COX-1 and COX-2 (Smith et al., 2000). COX-1, expressed constitutively in most cells, is considered the dominant, but not exclusive, source of prostanoids for housekeeping functions, such as cytoprotection of the gastric epithelium (see Chapter 45). COX-2, in contrast, is upregulated by cytokines, shear stress, and growth factors and is the principle source of prostanoid formation in inflammation and cancer. However, this distinction is overly simplistic; both enzymes contribute to the generation of autoregulatory and homeostatic prostanoids, and both can contribute to prostanoid formation during inflammation. Indeed, there are physiological and pathophysiological processes in which each enzyme is uniquely involved and others in which they function coordinately (Smith and Langenbach, 2001).

In addition to 61% amino acid identity, the crystal structures of COX-1 and COX-2 are remarkably similar (FitzGerald and Loll, 2001). Both isoforms are expressed as dimers homotypically inserted into the endoplasmic reticular membrane; their COX activity oxygenates and cyclizes unesterified AA to form prostaglandin G₂ (PGG₂), whereas their HOX activity converts PGG₂ to PGH₂ (Smith and Langenbach, 2001). These chemically unstable intermediates are transformed enzymatically into the prostanoids by isomerases and synthases. These enzymes are expressed in a relatively cell-specific fashion such that most cells make one or two dominant prostanoids. For example, COX-1-derived TxA₂ is the dominant product in platelets, whereas COX-2-derived PGE₂ and TxA₂ dominate in activated macrophages. Two classes of PGE synthases have been cloned. Microsomal (m) PGE synthases 1 and 2 co-localize with COX-2 in some, but not all, tissues and may be induced by cytokines and tumor promoters. Similarly, cytosolic PGE synthase often co-localizes with COX-1 and may be important in constitutive formation of PGE₂. Two forms of PGD synthase and PGF synthase have been identified. In heterologous expression systems, COX-1 couples preferentially with TxA₂ and PGF synthase, whereas COX-2 prefers PGI₂ synthase (Smyth and FitzGerald, 2009).

Prostanoids are released from cells predominantly by facilitated transport through the PG transporter and possibly other transporters (Schuster, 2002).

Products of Lipoxygenases. LOXs are a family of non-heme iron–containing enzymes that catalyze the oxygenation of polyenic fatty acids to corresponding lipid hydroperoxides (Brash, 1999). The enzymes require a fatty acid substrate with two cis double bonds separated by a methylene group. AA, which contains several double bonds in this configuration, is metabolized to hydroperoxy eicosatetraenoic acids (HPETEs), which vary in the site of insertion of the hydroperoxy group. Analogous to PGG₂ and PGH₂, these unstable intermediates, normally with S chirality, are further metabolized by a variety of enzymes. HPETEs are converted to their corresponding hydroxy fatty acid (HETE) either non-enzymatically or by a peroxidase.

There are five active human LOXs—5(S)-LOX, 12(S)-LOX, 12(R)-LOX, 15(S)-LOX-1, and 15(S)-LOX-2—classified according to the site of hydroperoxy group insertion. Human LOX products are of the S stereoconfiguration with the exception of 12(R)-LOX. Their expression is frequently cell specific (Brash, 1999); platelets have only 12(S)-LOX, whereas leukocytes contain both 5(S)- and 12(S)-LOX (Figure 33–2). 12(R)-LOX is restricted in expression mostly to the skin. The epidermal LOXs, which constitute a distinct LOX subgroup, also include 15-LOX-2 and eLOX-3, the most recently identified family member. eLOX-3 has been reported to metabolize further 12(R)-HETE, the product of 12(R)-LOX, to a specific epoxyalcohol product.

The 5-LOX pathway leads to the synthesis of the LTs, which play a major role in the development and persistence of the inflammatory response (Peters-Golden and Henderson, 2007) (Figure 33–2). A nomenclature (e.g., LTB₄, LTB₅) similar to that of prostanoids applies to the subclassification of the LTs. When eosinophils, mast cells, polymorphonuclear leukocytes, or monocytes are activated, 5-LOX translocates to the nuclear membrane and associates with 5-LOX-activating protein (FLAP), an integral membrane protein that facilitates AA to 5-LOX interaction (Evans et al., 2008). Drugs that inhibit FLAP block LT production. A two-step reaction is catalyzed by 5-LOX: oxygenation of AA at C5 to form 5-HPETE, followed by dehydration of 5-HPETE to an unstable 5,6-epoxide known as LTA₄. LTA₄ is transformed into bioactive eicosanoids by multiple pathways depending on the cellular context: transformation by LTA₄ hydrolase to a 5,12-dihydroxyeicosatetraenoic acid known as LTB₄; conjugation with GSH by LTC₄ synthase, in eosinophils, monocytes, and mast cells, to form LTC₄; and extracellular metabolism of the peptide moiety of LTC₄, leading to the removal of glutamic acid and subsequent cleavage of glycine, to generate LTD₄ and LTE₄, respectively (Peters-Golden and Henderson, 2007). LTC₄, LTD₄, and LTE₄, the cysteinyl leukotrienes (CysLTs), were known originally as the slow-reacting substance of anaphylaxis (SRS-A), first described >60 years ago. LTB₄ and LTC₄ are actively transported out of the cell.

At least two isoforms of 15-LOX exists, 15-LOX-1 and 15-LOX-2. The former prefers linoleic acid as a substrate and forms 15(S)-hydroxyoctadecadienoic acid, whereas the latter uses AA to generate 15(S)-HETE. Platelet-type 12(S)-LOX generates 12(S)-HETE from AA, whereas the leukocyte isozyme can synthesize both 12- and 15-HETE and often is referred to as 12/15-LOX. 12(S)-LOX can further metabolize LTA₄, the primary product of the 5-LOX pathway, to form the lipoxins LXA₄ and LXB₄. These mediators also can arise through 5-LOX metabolism of 15-HETE. 15(R)-HETE, derived from aspirin-acetylated COX-2, can be further transformed in leukocytes by 5-LOX to the epilipoxins 15-epi-LXA₄ or 15-epi-LXB₄, the so-called aspirin-triggered lipoxins (Brink et al., 2003). 12-HETE also can undergo a catalyzed molecular rearrangement to epoxyhydroxyeicosatrienoic acids called hepoxilins. Two additional groups of lipid mediators, the resolvins and protectins, may be generated from EPA and DHA, through the sequential action of aspirin-acetylated COX-2 and 5- or 15-LOX (Serhan et al., 2008).

The epidermal LOXs [12(R)-LOX, 15-LOX-2, and eLOX-3] are distinct from "conventional" enzymes in their substrate preferences and products (Furstenberger et al., 2007). Although the roles of epidermal LOXs in normal skin function are not clear, they may be relevant for skin barrier function and adipocyte differentiation. Epidermal accumulation of 12(R)-HETE is a feature of psoriasis and ichthyosis. Inhibitors of 12(R)-LOX are under investigation for the treatment of these proliferative skin disorders.

Products of CYPs. Multiple CYPs can metabolize AA (Capdevila and Falck, 2002). The epoxyeicosatrienoic acids (EETs), formed by CYP epoxygenases, primarily CYP2C and CYP2J in humans, have been a primary focus of research. Four regioisomers (14,15-; 11,12-; 8,9-; and 5,6-EETs), each containing a mixture of the (R,S) and (S,R) enantiomers, are formed in a CYP isoform–specific manner. EETs are synthesized in endothelial cells, where they function as endothelium-derived hyperpolarizing factors (EDHFs), particularly in the coronary circulation (Campbell and Falck, 2007). Their biosynthesis can be altered by pharmacological, nutritional, and genetic factors that affect CYP expression (see Chapter 6). CYP hydroxylases (primarily CYPs 4A, 4F) generate hydroxyeicosatetraenoic acids (16-, 17-, 18-, 19-, or 20-HETE), with 20-HETE being the principle product CYP-derived AA metabolite in vascular smooth muscle cells. 20-HETE is generated in response to smooth muscle cell stretch or to vasoactive agents, including angiotensin II (AngII).

EETs are metabolized by numerous pathways. The corresponding biologically less active (or inactive) dihydroxyeicosatrienoic acids (DHETs) are formed by epoxide hydrolases (EHs), whereas lysolipid acylation results in incorporation of EETs and DHETs into cellular phospholipids, where they can be stored. Inhibitors of EHs currently are under investigation. Glutathione conjugation and oxidation by COX and CYPs generate a series of glutathione conjugates, epoxyprostaglandins, diepoxides, tetrahydrofuran (THF) diols, and epoxyalcohols, whose biological relevance is not known. Similarly, 20-HETE can be converted by the COX pathway to the 20-hydroxy PGs. Intracellular fatty acid–binding proteins (FABPs) may bind EETs and DHETs differentially, thus modulating their metabolism, activities, and targeting.

Other Pathways. In addition to enzymatic formation of eicosanoids, several families of eicosanoid isomers are generated at significant concentrations in vivo by non-enzymatic free radical catalyzed oxidation of AA. The best characterized of these isoeicosanoids are the F₂-isoprostanes (F₂-IsoPs) (Lawson et al., 1999; Fam and Morrow, 2003; Milne et al., 2008). Unlike PGs, these compounds are initially formed esterified in phospholipids, after which they are hydrolyzed to their free form by phospholipases, including PAF acetylhydrolase (Stafforini et al., 2006), which then circulate and are metabolized and excreted into urine. Their production is not inhibited in vivo by inhibitors of COX-1 or COX-2, but their formation is suppressed by antioxidants. The PGF_2α isomer, 8-iso-PGF_2α, was the first F₂-IsoP to be identified. Measuring levels of these compounds in plasma and urine is considered the most accurate method to assess oxidative stress status in vivo, and increased levels are found in a large number of clinical conditions. Isoprostanes correlate with cardiovascular risk factors, but their use as predictors of coronary events remains an area of active investigation. Of particular interest is the recent finding that levels of F₂-IsoPs predict 30-day outcome in acute coronary syndrome (LeLeiko et al., 2009).

In addition to F₂-IsoPs, other PG-like isomers, including D₂/E₂-IsoPs, isothromboxanes, and isolevuglandins (alternatively termed isoketals), are also formed in vivo by non-enzymatic oxidation of AA (Brame et al., 2004). Isoleukotrienes are also generated non-enzymatically. Because several isoprostanes can activate prostanoid receptors, it has been speculated that they may contribute to the pathophysiology of inflammatory responses in a manner insensitive to COX inhibitors.

In the brain, the endocannabinoids arachidonylethanolamide (anandamide) and 2-arachidonoylglycerol are endogenous ligands of cannabinoid receptors (Bisogno, 2008). They mimic several pharmacological effects of Δ9-tetrahydrocannabinol, the active principle of Cannabis sativa preparations such as hashish and marijuana, including inhibition of adenylyl cyclase, inhibition of L-type Ca²⁺ channels, analgesia, and hypothermia. Several pathways have been proposed, but the prevalent one for biosynthesis in vivo remains unclarified. Conversion of anandamide and 2-arachidonylglycerol by COX-2 generates PG-ethanolamides (prostamides) and PG-glyceryl esters (Woodward et al., 2008); their biological significance remains to be clarified.

Compounds that preferentially and selectively inhibit the downstream enzymes that metabolize PGH₂ have been considered for some time. For example, agents that inhibit TxA₂ synthase might block platelet aggregation and induce vasodilation. Indeed, such drugs block TxA₂ production in vitro and in vivo; however, they have been disappointing in clinical development, perhaps owing to activation of the TxA₂ receptor by accumulated PGH₂ precursor. Given the problems associated with COX-2 inhibitors, there has been intense interest in developing drugs that might conserve the efficacy of COX-2 selectivity while shedding the cardiovascular risk. Microsomal prostaglandin E synthase-1 (mPGES-1), which catalyzes the isomerization of the COX product-PGH₂ into PGE₂, has emerged as a drug target (Samuelsson et al., 2007). The differential cardiovascular impact of mPGES-1 versus COX-2 inhibition in humans currently is unknown. However, deletion of mPGES-1 does not accelerate the response to a thrombogenic stimulus in vivo in rodents, in contrast to either selective inhibition of COX-2 or deletion of the IP, the I prostanoid receptor for PGI₂ (Cheng et al., 2006b). Indeed, deletion of mPGES-1 depresses systemic production of PGE₂ and augments biosynthesis of PGI₂ by rediversion of the intermediate COX product PGH₂ to PGI synthase. It currently is unclear what contribution elevated PGI₂ makes to the impact on mPGES-1 inhibition and whether this shift in substrate will be seen with pharmacological inhibitors in humans.

Because LTs mediate inflammation, efforts have focused on development of LT-receptor antagonists and selective inhibitors of the LOXs. Zileuton, an inhibitor of 5-LOX, and selective CysLT-receptor antagonists (zafirlukast, pranlukast, and montelukast) have established efficacy in the treatment of mild to moderate asthma (see Chapter 36). These treatments remain, however, less effective than inhaled corticosteroids. A common polymorphism in the gene for LTC₄ synthase that correlates with increased LTC₄ generation is associated with aspirin-intolerant asthma and with the efficacy of anti-LT therapy (Kanaoka and Boyce, 2004). Interestingly, although polymorphisms in the genes encoding 5-LOX or FLAP do not appear to be linked to asthma, studies have demonstrated an association of these genes with myocardial infarction, stroke, and atherosclerosis (Peters-Golden and Henderson, 2007); thus, inhibition of LT biosynthesis may be useful in the prevention of cardiovascular disease.

Eicosanoid Catabolism. Most eicosanoids are efficiently and rapidly inactivated. About 95% of infused PGE₂ (but not PGI₂) is inactivated during one passage through the pulmonary circulation. Broadly speaking, the enzymatic catabolic reactions are of two types: a relatively rapid initial step, catalyzed by widely distributed PG-specific enzymes, wherein PGs lose most of their biological activity, and a second step in which these metabolites are oxidized, probably by enzymes identical to those responsible for the β and ω oxidation of fatty acids (Figure 33–3). The initial step is the oxidation of the 15-OH group to the corresponding ketone by PG 15-OH dehydrogenase (PGDH) (Tai et al., 2002). Two types of 15-PGDHs have been identified. Type I, an NAD⁺-dependent enzyme, is the predominant form involved in eicosanoid catabolism. There is little circulating PGDH activity; thus, it is likely that metabolism first requires active transport to the intracellular space. The 15-keto compound then is reduced to the 13,14-dihydro derivative, a reaction catalyzed by PG Δ¹³-reductase. This enzyme is identical to the LTB₄ 12-hydroxydehydrogenase (see the discussion later). Subsequent steps consist of β and ω oxidation of the PG side chains, giving rise to polar dicarboxylic acids in the case of PGEs, which then are excreted in the urine as major metabolites (Figure 33–1); these reactions occur primarily in the liver.

Unlike PGE₂, PGD₂ initially is reduced in vivo to the F-ring PG, 9α11β-PGF₂, which possesses significant biological activity. Subsequently, this compound undergoes metabolism similar to that of other eicosanoids (Figure 33–3). TxA₂ breaks down non-enzymatically (t_1/2 = 30 seconds) into the chemically stable but biologically inactive thromboxane B₂ (TxB₂), which then is further metabolized by 11-hydroxy TxB₂ dehydrogenase to generate 11-dehydro-TxB₂ or by β oxidation to form 2,3-dinor-TxB₂ (Figure 33–3).

The degradation of PGI₂ (t_1/2 = 3 minutes) apparently begins with its spontaneous hydrolysis in blood to 6-keto-PGF_1α. The metabolism of this compound in humans involves the same steps as those for PGE₂ and PGF_2α.

The degradation of LTC₄ occurs in the lungs, kidney, and liver. The initial steps involve its conversion to LTE₄. LTC₄ also may be inactivated by oxidation of its cysteinyl sulfur to a sulfoxide. In leukocytes, LTB₄ is inactivated principally by oxidation by members of the CYP4F subfamily. Conversion to 12-oxo-LTB₄ by LTB 12-OH dehydrogenase (see the discussion earlier) is a key pathway in tissues other than leukocytes.

Prostaglandin Receptors. PGs activate membrane receptors locally near their sites of formation. The diversity of their effects is explained to a large extent by their interaction with a diverse family of distinct receptors (Table 33–1). Eicosanoid receptors interact with G_s, G_i, and G_q to modulate the activities of adenylyl cyclase and phospholipase C (see Chapter 3). Single gene products have been identified for the receptors for PGI₂ (the IP), PGF_2α (the FP), and TxA₂ (the TP). Four distinct PGE₂ receptors (EP_1-4) and two PGD₂ receptors (DP₁ and DP₂—also known as CRTH₂) have been cloned. Additional isoforms of the TP (α and β), FP (A and B), and EP₃ (I-VI, e, f) receptors can arise through differential mRNA splicing (Narumiya et al., 1999; Smyth et al., 2009).

Cell Signaling Pathways and Expression. The prostanoid receptors appear to derive from an ancestral EP receptor and share high homology. Phylogenetic comparison of this family reveals three subclusters: The first consists of the relaxant receptors EP₂, EP₄, IP, and DP₁, which increase cellular cyclic AMP generation; the second consists of the contractile receptors EP₁, FP, and TP, which increase cytosolic levels of Ca²⁺; and a third, consisting only of EP₃, can couple to both elevation of intracellular calcium and a decrease in cyclic AMP. The DP₂ receptor is an exception and is unrelated to the other prostanoid receptors; rather, it is a member of the formyl-methionyl-leucyl-phenylalanine (fMLP)-receptor superfamily (Table 33–1).

TP_α and TP_β receptors couple via G_q and several other G proteins, to activate the PLC–IP₃–Ca²⁺ pathway (Figure 33–4). Activation of TP receptors also may activate or inhibit adenylyl cyclase via G_s (TP_α) or G_i (TP_β), respectively, and signal via G_q, G_12/13, and G₁₆, to stimulate small G protein–signaling pathways, including ERK and Rho. TP is expressed in platelets, vasculature, lung, kidney, heart, thymus, and spleen. The TP_α apparently is the sole isoform expressed in platelets (see Smyth and FitzGerald, 2009). Recognized differences between the splice variants are limited to G protein activation and receptor regulation in heterologous in vitro expression systems.

IP couples with G_s to stimulate adenylyl cyclase activity. It is expressed in many tissues and cells, including human kidney, lung, spine, liver, vasculature, and heart.

DP₁ also couples with adenylyl cyclase through G_s. It is the least abundant of the prostanoid receptors, with expression in mouse ileum, lung, stomach, and uterus. In humans, several subtypes of leukocytes express DP₁, including eosinophils, basophils, monocytes, dendritic cells, and T lymphocytes. DP₁ also is expressed in the central nervous system (CNS), where it appears to be limited specifically to the leptomeninges and the vasculature. DP₂ couples with the G_q–PLC–IP₃ pathway to increase intracellular Ca²⁺ (Figure 33–4). Its mRNA is found in many human tissues (brain, heart, thymus, spleen, liver, and intestine). DP₂ is expressed on mast cells, T lymphocytes (Th2 but not Th1), basophils, and eosinophils (Kim and Luster, 2007).

EP₂ and EP₄ receptors activate adenylate cyclase via G_s. The EP₂ receptor is expressed at much lower levels in most tissues and can be induced in response to inflammatory stimuli, suggesting distinct roles for these two G_s-coupled EP receptors. The EP₁ receptor, via an unclassified G protein, can activate the PLC–IP₃–Ca²⁺ pathway. All human EP₃ isoforms can activate G_i to inhibit adenylyl cyclase; however, certain isoforms can activate G_s or G_12/13. EP₁ and EP₂ receptors have limited distribution compared with the distribution of EP₃ and EP₄ receptors.

FP_A and FP_B receptors couple via G_q–PLC–IP₃ to mobilize cellular Ca²⁺ and activate PKC. In addition, stimulation of FP activates Rho kinase, leading to the formation of actin stress fibers, phosphorylation of p125 focal adhesion kinase, and cell rounding. The FP receptor is expressed in kidney, heart, lung, stomach, and eye; it is most abundant in the corpus luteum, where its expression pattern varies during the estrus cycle.

Leukotriene and Lipoxin Receptors. Several receptors for the LTs and lipoxins have been identified (Peters-Golden and Henderson, 2007) (Table 33–1). Two receptors exist for both LTB₄ (BLT₁ and BLT₂) and the cysteinyl leukotrienes (CysLT₁ and CysLT₂). A receptor that binds lipoxin, ALX, is identical to the fMLP-1 receptor; the nomenclature now reflects LXA₄ as a natural and potent ligand (Chiang et al., 2006).

Cell Signaling Pathways and Expression. Phylogenetic comparison reveals two clusters of LT/lipoxin receptors: the chemoattractant receptors (BLT₁, BLT₂, and ALX), which also contain the DP₂ receptor for PGD₂, and the CysLT₁ and CysLT₂ receptors. All are G protein–coupled receptors (GPCRs) and couple with G_q and other G proteins (Table 33–1), depending on the cellular context. The BLT₁ is expressed predominantly in leukocytes, thymus, and spleen, whereas BLT₂, the low-affinity receptor for LTB₄, is found in spleen, leukocytes, ovary, liver, and intestine. BLT₂ binds 12(S)- and 12(R)-HETE with reasonable affinity, although the biological relevance of this observation is not clear.

CysLT₁ binds LTD₄ with higher affinity than LTC₄, while CysLT₂ shows equal affinity for both LTs. Both receptors bind LTE₄ with low affinity. Activation of G_q, leading to increased intracellular Ca²⁺, is the primary signaling pathway reported. Studies also have placed G_i downstream of CysLT₂. CysLT₁ is expressed in lung and intestinal smooth muscle, spleen, and peripheral blood leukocytes, whereas CysLT₂ is found in heart, spleen, peripheral blood leukocytes, adrenal medulla, and brain.

Responses to ALX-receptor activation vary with cell type. In human neutrophils, AA release is stimulated, whereas Ca²⁺ mobilization is blocked; in monocytes, LXA₄ stimulates Ca²⁺ mobilization. The ALX receptor is expressed in lung, peripheral blood leukocytes, and spleen.

Other Agents. Other AA metabolites (e.g., isoprostanes, epoxy-eicosatrienoic acids, hepoxilins) have potent biological activities, and there is evidence for distinct receptors for some of these substances. Some isoprostanes appear to act as incidental ligands at the TP (Audoly et al., 2000), which may be important in the pathology of cardiovascular disease. Others activate the FP (Kunapuli et al., 1997). Certain eicosanoids, most notably 15-deoxy-Δ^12,14-PGJ₂ (15d-PGJ₂), a dehydration product of PGD₂, have been reported as endogenous ligands for a family of nuclear receptors called peroxisome proliferator–activated receptors (PPARs) that regulate lipid metabolism and cellular proliferation and differentiation. However, their affinities for PPARs are significantly lower than for cell surface receptors, raising doubt about the physiological relevance of the ligand–receptor interaction. in vitro 15d-PGJ₂ can bind PPARγ, but the quantities formed in vivo are orders of magnitude lower than those necessary for PPAR activation (Bell-Parikh et al., 2003). Specific receptors for the HETEs and EETs have been proposed but not yet isolated.

Platelets. Platelet aggregation leads to activation of membrane phospholipases, with the release of AA and consequent eicosanoid biosynthesis. In human platelets, TxA₂ and 12-HETE are the two major eicosanoids formed, although eicosanoids from other sources (e.g., PGI₂ derived from vascular endothelium) also affect platelet function. A naturally occurring mutation in the first intracellular loop of the TP receptor is associated with a mild bleeding diathesis and resistance of platelet aggregability to TP agonists (Hirata et al., 1994). The importance of the TxA₂ pathway is evident from the efficacy of low-dose aspirin in the secondary prevention of myocardial infarction and ischemic stroke. The total biosynthesis of TxA₂, as determined by excretion of its urinary metabolites, is augmented in clinical syndromes of platelet activation, including unstable angina, myocardial infarction, and stroke (Smyth et al., 2009). Deletion of the TP receptor in the mouse prolongs bleeding time, renders platelets unresponsive to TP agonists, modifies their response to collagen but not to ADP, and blunts the response to vasopressors and the proliferative response to vascular injury.

PGI₂ inhibits platelet aggregation and disaggregates preformed clumps. Deficiency of the IP receptor in disease-free mice does not alter platelet aggregation significantly ex vivo, although increased responsiveness to thrombin was evident in a mouse model of atherosclerosis (Smyth and FitzGerald, 2009). Augmented biosynthesis of PGI₂ in syndromes of platelet activation serves to constrain the effects platelet agonists, vasoconstrictors, and stimuli to platelet activation. However, PGI₂ does limit platelet activation by TxA₂ in vivo, reducing the thrombotic response to vascular injury (Cheng et al., 2002). The increased incidence of myocardial infarction and stroke in patients receiving selective inhibitors of COX-2, most parsimoniously explained by inhibition of COX-2-dependent PGI₂ formation, supports this concept (Grosser et al., 2006b).

Low concentrations of PGE₂ activate the EP₃ receptor, leading to platelet aggregation (Fabre, 2001). Deletion of the EP₃ in mice leads to an increased bleeding tendency and decreased susceptibility to thromboembolism. Deletion of mPGES-1 did not affect thrombogenesis in vivo, probably due to substrate rediversion and augmented formation of PGI₂ (Cheng et al., 2006b).

Vascular Tone. Because of their short t_1/2, prostanoids do not circulate and generally are considered not to impact directly on systemic vascular tone. They may, however, modulate vascular tone locally at their sites of biosynthesis or through renal or other indirect effects. PGI₂, the major arachidonate metabolite released from the vascular endothelium, is derived primarily from COX-2 in humans (Catella-Lawson et al., 1999; McAdam et al., 1999). PGI₂ generation and release is regulated by shear stress and by both vasoconstrictor and vasodilator autacoids. Deletion of the IP in mice augments vascular proliferation, remodeling, atherogenesis, and hypertension, while PGI synthase polymorphisms have been associated with essential hypertension and myocardial infarction (Smyth and FitzGerald, 2009). PGI₂ limits pulmonary hypertension induced by hypoxia and systemic hypertension induced by AngII and lowers pulmonary resistance in patients with pulmonary hypertension.

Deficiency of EP₁ or EP₄ receptors reduces resting blood pressure in male mice; EP₁-receptor deficiency is associated with elevated renin–angiotensin activity. Both EP₂ and EP₄ receptor–deficient animals develop hypertension in response to a high-salt diet, reflecting the importance of PGE₂ in maintenance of renal blood flow and salt excretion. PGI₂ and PGE₂ are implicated in the hypotension associated with septic shock. PGs also may play a role in the maintenance of placental blood flow. Although conflicting data exist, it appears that loss of mPGES-1 in mice is less likely than loss of COX-2 to alter blood pressure (Smyth et al., 2009).

COX-2-derived PGE₂, via the EP₄ receptor, maintains the ductus arteriosus patent until birth, when reduced PGE₂ levels (a consequence of increased PGE₂ metabolism) permit closure. (Coggins et al., 2002). The tNSAIDs induce closure of a patent ductus in neonates (see Chapter 34). Contrary to expectation, animals lacking the EP₄ receptor die with a patent ductus during the perinatal period (Table 33–1) because the mechanism for control of the ductus in utero, and its remodeling at birth, is absent.

Endogenous biosynthesis of EETs is increased in human syndromes of hypertension. An analog of 11,12-EET abrogated the enhanced renal microvascular reactivity to AngII associated with hypertension (Imig et al., 2001), and blood pressure is lower in mice deficient in soluble EH (Sinal et al., 2000); these findings suggest that EH enzyme may be a potential pharmacological target for hypertension. Much indirect evidence suggests the existence of EET receptors, although none has been cloned.

Inflammatory Vascular Disease. Studies with knockout mice strongly implicate prostanoids in the development of atherogenesis and abdominal aortic aneurism; both inflammatory cardiovascular diseases (Smyth et al., 2009; Smyth and FitzGerald, 2009). Suppression of TxA₂ biosynthesis, as well as antagonism or deletion of the TP, retards atherogenesis in mice. Deletion of the FP receptor reduces blood pressure and retards atherogenesis, coincident with reduced renin. Conversely, PGI₂ appears atheroprotective and also limits vascular proliferative and remodeling responses. In humans, an arginine²¹² to cysteine substitution in the fifth intracellular loop of the IP, which disrupts IP signaling, co-segregated with increased cardiovascular risk in a recent study (Arehart et al., 2008), concordant with a role for this prostanoid in modifying human cardiovascular disease.

The role of PGE₂ effects on inflammatory cardiovascular disease is less clear. Deletion of mPGES-1 does not accelerate the response to a thrombogenic stimulus in vivo in rodents (in contrast to either selective inhibition of COX-2 or deletion of the IP (Cheng et al., 2006b) but does retard atherogenesis in fat-fed hyperlipidemic mice (Wang et al., 2006). It is unclear, however, whether this results from loss of PGE₂ or because of concomitant elevations in PGI₂ biosynthesis. Deletion, or selective inhibition, of COX-2, but not inhibition of COX-1, decreases abdominal aortic aneurism formation in hyperlipidemic mice (King et al., 2006). Similar results were seen in mPGES-1-deficient mice (Wang et al., 2008), although, again, it is unclear to what extent rediversion to biosynthesis of other prostanoids (e.g., PGI₂) contributes.

There is growing evidence for a role of the LTs in cardiovascular disease (Peters-Golden and Henderson, 2007). Although conflicting data have been reported in animal studies, human genetic studies have demonstrated a link between cardiovascular disease and polymorphisms in the LT biosynthetic enzymes and FLAP.

Lung. A complex mixture of autacoids is released when sensitized lung tissue is challenged by the appropriate antigen. COX-derived bronchodilator (PGE₂) and bronchoconstrictor (e.g., PGF_2α, TxA₂, PGD₂) substances are released. IP deletion in mice exaggerates features of acute and chronic experimental asthma, including increased bronchial hyperresponsiveness. Inhaled iloprost (a PGI₂ analog) suppresses the cardinal features of asthma in mice via inhibition of airway dendritic cell function.

Polymorphisms in the genes for PGD₂ synthase and the TP receptor have been associated with asthma in humans. Deletion of either DP₁ or DP₂ in mice suggests an important role of this prostanoid in asthma (and in other allergic responses), although contradictory findings in DP₂-deficient mice suggest significant complexity in the function of PGD₂ in airway inflammation (Pettipher et al., 2007).

The CysLTs probably dominate during allergic constriction of the airway (Drazen, 1999). Deficiency of 5-LOX leads to reduced influx of eosinophils in airways and attenuates bronchoconstriction. Furthermore, unlike COX inhibitors and histaminergic antagonists, CysLT-receptor antagonists and 5-LOX inhibitors are effective in the treatment of human asthma (see "Inhibitors of Eicosanoid Biosynthesis"). The relatively slow LT metabolism in lung contributes to the long-lasting bronchoconstriction that follows challenge with antigen and may be a factor in the high bronchial tone that is observed in asthmatic patients in periods between acute attacks (see Chapter 36).

Kidney. Long-term use of all COX inhibitors is limited by the development of hypertension, edema, and congestive heart failure in a significant number of patients. PGE₂, along with PGI₂, apparently derived from COX-2, plays a critical role in maintaining renal blood flow and salt excretion, whereas there is some evidence that the COX-1-derived vasoconstrictor TxA₂ may play a counterbalancing role. Biosynthesis of PGE₂ and PGI₂ is increased by factors that reduce renal blood flow (e.g., stimulation of sympathetic nerves; AngII).

Bartter's syndrome is an autosomal recessive trait that is manifested as hypokalemic metabolic alkalosis. The syndrome results from inappropriate renal salt absorption caused primarily by dysfunctional mutations in the Na⁺–K⁺–2Cl^– co-transporter NKCC2, a target of loop diuretics in the ascending thick limb of the loop of Henle (Simon et al., 1996) (see Chapter 25). The syndrome also can result from dysfunctional alterations in proteins whose activities can limit NKCC2 function: the K⁺ channel ROMK2 (Kir1.1) that recycles K⁺ into the tubular fluid; the basolateral membrane Cl^– channel, ClC–Kb; and Barttin, the integral membrane protein that forms the α-subunit of the ClC–Kb heteromer (O'Shaughnessy and Karet, 2004). The antenatal variant of Bartter's syndrome, owing to dysfunctional ROMK2, also is known as hyperprostaglandin E syndrome. The elevated PGE₂ may exacerbate the symptoms of salt and water loss. The relationship between dysfunctional ROMK2 and elevated PGE₂ synthesis is not clear. However, in patients with antenatal Bartter's syndrome, inhibition of COX-2 ameliorates many of the clinical symptoms (Nüsing et al., 2001).

Inflammatory and Immune Responses. PGs and LTs are synthesized in response to a host of stimuli that elicit inflammatory and immune responses, and contribute significantly to inflammation and immunity (Tilley et al., 2001; Brink et al., 2003; Kim and Luster, 2007). Prostanoids generally promote acute inflammation, although there are some exceptions, such as the inhibitory actions of PGE₂ on mast cell activation (see "Inflammation and Immunity"). Data from animals deficient in either COX-1 or COX-2 yield conflicting results depending on the inflammatory model used, perhaps reflecting the contribution of both isozymes to inflammation. Deletion of mPGES markedly reduced inflammation in several mouse models.

LTs are potent mediators of inflammation. Deletion of either 5-LOX or FLAP reduces inflammatory responses. Generation of BLT₁-deficient mice confirms the role of LTB₄ in chemotaxis, adhesion, and recruitment of leukocytes to inflamed tissues (Toda et al., 2002). Increased vascular permeability resulting from innate and adaptive immune challenges is offset in mice deficient in CysLT₁ or LTC₄ synthase (Kanaoka and Boyce, 2004) (Table 33–1). Deletion either of LTC₄ synthase (and thus loss of CysLT biosynthesis) or CysLT₂ reduced chronic pulmonary inflammation and fibrosis in response to bleomycin. In contrast, absence of CysLT₁ led to an exaggerated response. These findings demonstrate a role for CysLT₂ in promoting, and an unexpected role for CysLT₁ in counteracting, chronic inflammation.

Heart. Studies suggest a role for COX-2 in cardiac function (Smyth et al., 2009). PGI₂ and PGE₂, acting on the IP or the EP₃, respectively (Dowd et al., 2001; Shinmura et al., 2005), protect against oxidative injury in cardiac tissue. IP deletion augments myocardial ischemia/reperfusion injury and both mPGES-1 deletion (Degousee et al., 2008) and cardiomyocyte specific deletion of the EP₄ (Qian et al., 2008) exacerbate the decline in cardiac function after experimental myocardial infarction. COX-2-derived TxA₂ contributed to oxidant stress, isoprostane generation, and activation of the TP, and also possibly the FP, to increase cardiomyocyte apoptosis and fibrosis in a model of heart failure (Zhang et al., 2003). Selective deletion of COX-2 in cardiomyocytes results in mild heart failure and a predisposition to arrhythmogenesis.

Reproduction and Parturition. Studies with knockout mice confirm a role for PGs in reproduction and parturition (Smyth and FitzGerald, 2009). COX-1-derived PGF_2α appears important for luteolysis, consistent with delayed parturition in mice deficient in COX-1. Subsequent upregulation of COX-2 generates prostanoids, including PGF_2α and TxA₂, that are important in the final stages of parturition. Mice lacking both COX-1 and oxytocin undergo normal parturition, demonstrating the critical interplay between PGF_2α and oxytocin in onset of labor. EP₂ receptor–deficient mice demonstrate a preimplantation defect (Table 33–1), which likely underlies some of the breeding difficulties seen in COX-2 knockouts.

Cancer. Pharmacological inhibition or genetic deletion of COX-2 restrains tumor formation in models of colon, breast, lung, and other cancers. Large human epidemiological studies report that the incidental use of NSAIDs is associated with significant reductions in relative risk for developing these and other cancers (Harris et al., 2005). PGE₂ has been implicated as the primary pro-oncogenic prostanoid in multiple studies. The pro- and anti-oncogenic roles of other prostanoids remain under investigation, with TxA₂ emerging as another likely COX-2-derived pro-carcinogenic mediator. Studies in mice lacking EP₁, EP₂, and EP₄ reduce disease in multiple carcinogenesis models. EP₃, in contrast, may even play a protective role in some cancers. Three randomized controlled trials of COX-2 inhibitors reported a significant reduction in the reoccurrence of adenomas in patients receiving either celecoxib or rofecoxib compared to placebo (Bertagnolli, 2007), while polymorphisms in COX-2 have been associated with an increased risk of colon and other cancers. In mouse mammary tissue, COX-2 is pro-oncogenic (Liu et al., 2001), whereas aspirin use is associated with a reduced risk of breast cancer in women, especially for hormone receptor–positive tumors (Terry et al., 2004). Despite the emphasis on COX-2, studies support a role for both COX enzymes in pro-oncogenic processes, and it remains untested whether selective COX-2 inhibitors will prove superior to nonselective NSAIDs for the prevention or treatment of human cancer. The CysLT and LTB₄ receptors also are implicated in cancer, raising interest in the use of LT inhibitors/antagonists in chemoprevention/therapy.

Cardiac output generally is increased by infusion of PGs of the E and F series. Weak, direct inotropic effects have been noted in various isolated preparations. In the intact animal, however, increased force of contraction and increased heart rate are, in large measure, a reflex consequence of a fall in total peripheral resistance. Animal studies suggest direct cardioprotective effects of PGI₂ and PGE₂ (Smyth et al., 2009).

LTC₄ and LTD₄ can constrict or relax isolated vascular smooth muscle preparations, depending on the concentrations used and the vascular bed (Brink et al., 2003). Hypotension in humans may result partly from a decrease in intravascular volume and also from decreased cardiac contractility secondary to a marked LT-induced reduction in coronary blood flow. Although LTC₄ and LTD₄ have little effect on most large arteries or veins, coronary arteries and distal segments of the pulmonary artery are contracted by nanomolar concentrations of these agents. The renal vasculature is resistant to this constrictor action, but the mesenteric vasculature is not. LTC₄ and LTD₄ act in the microvasculature to increase permeability of postcapillary venules; they are approximately 1,000-fold more potent than histamine in this regard. At higher concentrations, LTC₄ and LTD₄ can constrict arterioles and reduce exudation of plasma. Vascular smooth muscle proliferation can be promoted by 12S-HETE and 20-HETE.

EETs cause vasodilation in a number of vascular beds by activating the large conductance Ca²⁺-activated K⁺ channels of smooth muscle cells, thereby hyperpolarizing the smooth muscle and causing relaxation. EETs likely also function as endothelium-derived hyperpolarizing factors (EDHFs), particularly in the coronary circulation (Campbell and Falck, 2007). In contrast to EETs, 20-HETE inhibits large conductance Ca²⁺-activated K⁺ channels, resulting in depolarization of the vascular smooth muscle cell, Ca²⁺ entry, and potent vasoconstriction (Kroetz and Xu, 2005). Evidence supports the role of 20-HETE in the regulation of vascular tone, particularly in renal autoregulation.

Isoprostanes usually are vasoconstrictors, although there are examples of vasodilation in preconstricted vessels.

Mature platelets express only COX-1. Megakaryocytes and immature platelet forms, released in clinical conditions of accelerated platelet turnover, also express COX-2 (Rocca et al., 2002), but its role in platelet development and function has yet to be elucidated. TxA₂, the major product of COX-1 in platelets, induces platelet shape change, through G₁₂/G₁₃-mediated Rho/Rho-kinase-dependentregulation of myosin light-chain phosphorylation, and aggregation through G_q-dependent activation of PKC. Perhaps more importantly, TxA₂ amplifies the signal for other, more potent platelet agonists, such as thrombin and ADP (FitzGerald, 1991). The actions of TxA₂ on platelets are restrained by its short t_1/2 (∼30 seconds), by rapid TP desensitization, and by endogenous inhibitors of platelet function, including NO and PGI₂, which inhibits platelet aggregation by all recognized agonists. The biological importance of 12-HETE formation is poorly understood, although deletion of the platelet 12-LOX augments ADP-induced platelet aggregation and AA-induced sudden death in mice. Some isoprostanes increase the response of platelets to pro-aggregatory agonists in vitro.

COX-2 is the major source of prostanoids formed during and after an inflammatory response, although COX-1 also contributes. PGE₂ and PGI₂ are the predominant pro-inflammatory prostanoids, as a result of increased vascular permeability and blood flow in the inflamed region. TxA₂ can increase platelet–leukocyte interaction. Prostanoids, especially PGD₂, also contribute to resolution of inflammation. PGs generally inhibit lymphocyte function and proliferation, suppressing the immune response (Rocca et al., 2002). PGE₂ depresses the humoral antibody response by inhibiting the differentiation of B lymphocytes into antibody-secreting plasma cells. PGE₂ acts on T lymphocytes to inhibit mitogen-stimulated proliferation and lymphokine release by sensitized cells. PGE₂ and TxA₂ also may play a role in T lymphocyte development by regulating apoptosis of immature thymocytes (Tilley et al., 2001). PGD₂, a major product of mast cells, is a potent leukocyte chemoattractant (Pettipher et al., 2007), primarily through the DP₂. Activation of the DP₂ promotes chemotaxis and activation of T_H2 lymphocytes, eosinophils, and basophils. The PGD₂ degradation product, 15d-PGJ₂, also may activate eosinophils via the DP₂. PDG₂-mediated polarization of T cells to the T_H2 phenotype is DP₁-mediated. A counterregulatory role for DP₁ in leukocyte activation has been proposed, although complementary functions for the two receptors have been reported. The precise interplay between DP₁ and DP₂ in vivo remains to be clarified.

LTB₄ is a potent activator and chemotactic agent for neutrophils, T lymphocytes, eosinophils, monocytes, dendritic cells, and possibly also mast cells (Kim and Luster, 2007). These effects are primarily BLT₁ receptor-mediated and, although BLT₂ expression has been reported in eosinophils, mast cells, and dendritic cells, its contribution to LBT₄ function is unclear. LTB₄ stimulates the aggregation of eosinophils and promotes degranulation and the generation of superoxide. LTB₄ promotes adhesion of neutrophils to vascular endothelial cells and their transendothelial migration and stimulates synthesis of pro-inflammatory cytokines from macrophages and lymphocytes. Mast cell–generated LTB₄ also may contribute significantly to T lymphocyte migration.

The CysLTs are chemotaxins for eosinophils and monocytes through activation of the CysLT₁ receptor. They also induce cytokine generation in eosinophils, mast cells, and dendritic cells. A distinct set of cytokines are produced through activation of mast cell CysLT₁ and CysLT₂. At higher concentrations, these LTs also promote eosinophil adherence, degranulation, cytokine or chemokine release, and oxygen radical formation. In addition, CysLTs contribute to inflammation by increasing endothelial permeability, thus promoting migration of inflammatory cells to the site of inflammation.

Lipoxins have diverse effects on leukocytes, including activation of monocytes and macrophages and inhibition of neutrophil, eosinophil, and lymphocyte activation (McMahon and Godson, 2004). Both lipoxin A and B inhibit natural killer cell cytotoxicity.

Bronchial and Tracheal Muscle. In general, TxA₂, PGF_2α, and PGD₂ contract, and PGE₂ and PGI₂ relax, bronchial and tracheal muscle. PGD₂ appear to be the primary bronchoconstrictor prostanoid of relevance in humans. Roughly 10% of people given aspirin or tNSAIDs develop bronchospasm. This appears attributable to a shift in AA metabolism to LT formation, as reflected by an increase in urinary LTE₄ in response to aspirin challenge in such individuals. This substrate diversion appears to involve COX-1; such patients do not develop bronchospasm when treated with selective inhibitors of COX-2. The CysLTs are bronchoconstrictors in many species, including humans (Brink et al., 2003). These LTs act principally on smooth muscle in the airways and are a thousand times more potent than histamine both in vitro and in vivo. They also stimulate bronchial mucus secretion and cause mucosal edema.

PGI₂ causes bronchodilation in most species; human bronchial tissue is particularly sensitive, and PGI₂ antagonizes bronchoconstriction induced by other agents.

Uterus. Strips of nonpregnant human uterus are contracted by PGF_2α and TxA₂ but are relaxed by PGEs. Sensitivity to the contractile response is most prominent before menstruation, whereas relaxation is greatest at midcycle. Uterine strips obtained at hysterectomy from pregnant women are contracted by PGF_2α and by low concentrations of PGE₂. PGE₂, together with oxytocin, is essential for the onset of parturition. PGI₂ and high concentrations of PGE₂ produce relaxation. The intravenous infusion of PGE₂ or PGF_2α to pregnant women produces a dose-dependent increase in uterine tone and in the frequency and intensity of rhythmic uterine contractions. PGEs and PGFs are used to terminate pregnancy. Uterine responsiveness to PGs increases as pregnancy progresses but remains smaller than the response to oxytocin.

Gastrointestinal Muscle. PGEs and PGFs stimulate contraction of the main longitudinal muscle from stomach to colon. PG endoperoxides, TxA₂, and PGI₂ also produce contraction but are less active. Circular muscle generally relaxes in response to PGE₂ and contracts in response to PGF_2α. The LTs have potent contractile effects. PGs reduce transit time in the small intestine and colon. Diarrhea, cramps, and reflux of bile have been noted in response to oral PGE; these are common side effects (along with nausea and vomiting) in patients given PGs for abortion. PGEs and PGFs stimulate the movement of water and electrolytes into the intestinal lumen. Such effects may underlie the watery diarrhea that follows their oral or parenteral administration. By contrast, PGI₂ does not induce diarrhea; indeed, it prevents that provoked by other PGs.

PGE₂ appears to contribute to the water and electrolyte loss in cholera, a disease that is somewhat responsive to therapy with tNSAIDs.

TxA₂, generated at low levels in the normal kidney, has potent vasoconstrictor effects that reduce renal blood flow and glomerular filtration rate. Infusion of PGF_2α causes both natriuresis and diuresis. Conversely, PGF_2α may activate the renin–angiotensin system, contributing to elevated blood pressure.

There is substantial evidence for a role of the CYP epoxygenase products in regulating renal function, although their exact role in the human kidney remains unclear. Both 20-HETE and the EETs are generated in renal tissue. 20-HETE constricts the renal arteries, while EETs mediate vasodilation and natriuresis.

COX-2-derived prostanoids also have been implicated in several CNS degenerative disorders (e.g, Alzheimer's disease, Parkinson disease; see Chapter 22), although the therapeutic efficacy of blocking their synthesis or action remains to be established.

LOX metabolites also have endocrine effects. 12-HETE stimulates the release of aldosterone from the adrenal cortex and mediates a portion of the aldosterone release stimulated by AngII, but not that which occurs in response to ACTH.

Therapeutic Abortion. The ability of PGEs, PGFs, and their analogs, to terminate pregnancy at any stage by promoting uterine contractions has been adapted to common clinical use. When given early in pregnancy, their action as abortifacients may be variable and often incomplete and accompanied by adverse effects. PGs appear, however, to be of value in missed abortion and molar gestation, and they have been used widely for the induction of midtrimester abortion. Dinoprostone, a synthetic preparation of PGE₂, is approved for inducing abortion in the second trimester of pregnancy, for missed abortion, for cervical ripening prior to induction of labor, and for managing benign hydatidiform moles. Several studies have shown that systemic or intravaginal administration of the PGE₁ analog misoprostol in combination with mifepristone (RU486) or methotrexate is highly effective in the termination of early pregnancy. An analog of PGF_2α, carboprost tromethamine, is used to induce second-trimester abortions and to control postpartum hemorrhage that is not responding to conventional methods.

PGE₂ or PGF_2α can induce labor at term. However, they may have more value when used to facilitate labor by promoting ripening and dilation of the cervix.

Gastric Cytoprotection. The capacity of several PG analogs to suppress gastric ulceration is a property of therapeutic importance. Misoprostol (cytotec), a PGE₁ analog, is approved for prevention of NSAID-induced gastric ulcers. It is about as effective as the proton pump inhibitor omeprazole (Chapter 45).

Impotence. PGE₁ (alprostadil), given as an intracavernous injection (caverject, edex) or urethral suppository (muse) is a second-line treatment of erectile dysfunction. The erection lasts for 1-3 hours and is sufficient for sexual intercourse. PGE₁ has been superseded largely by the use of phosphodiesterase 5 (PDE5) inhibitors, such as sildenafil, tadalafil, and vardenafil (see Chapters 27 and 28).

Maintenance of Patent Ductus Arteriosus. The ductus arteriosus in neonates is highly sensitive to vasodilation by PGE₁. Maintenance of a patent ductus may be important hemodynamically in some neonates with congenital heart disease. PGE₁ (alprostadil, PROSTIN VR PEDIATRIC) is highly effective for palliative, but not definitive, therapy to maintain temporary patency until surgery can be performed. Apnea is observed in ∼10% of neonates treated, particularly those who weigh <2 kg at birth.

Pulmonary Hypertension. Primary pulmonary hypertension is a rare idiopathic disease that mainly affects young adults. It leads to right-sided heart failure and frequently is fatal. Long-term therapy with PGI₂ (prostacyclin; epoprostenol, FLOLAN), via continuous intravenous infusion, improves symptoms and can delay or preclude the need for lung or heart-lung transplantation in a number of patients. Epoprostenol also has been used successfully to treat portopulmonary hypertension that arises secondary to liver disease, again with a goal of facilitating ultimate transplantation (Krowka et al., 1999). Several PGI₂ analogs with longer t_1/2 have been developed and used clinically. Iloprost can be inhaled (ventavis) or delivered by intravenous administration (not available in the U.S.). Treprostinil (remoduln) (t_1/2 ∼4 hours) may be delivered by continuous subcutaneous or intravenous infusion.

Glaucoma. Latanoprost, a stable, long-acting PGF_2α derivative, was the first prostanoid used for glaucoma. The success of latanoprost has stimulated development of similar prostanoids with ocular hypotensive effects, and bimatoprost and travoprost are now available. These drugs act as agonists at the FP receptor and are administered as ophthalmic drops (Chapter 64).

History. In 1971, Henson demonstrated that a soluble factor released from leukocytes caused platelets to aggregate. Benveniste and his coworkers characterized the factor as a polar lipid and named it platelet-activating factor. During this period, Muirhead described an antihypertensive polar renal lipid (APRL) produced by interstitial cells of the renal medulla that proved to be identical to PAF. Hanahan and coworkers then synthesized acetyl glyceryl ether phosphorylcholine (AGEPC) and determined that this phospholipid had chemical and biological properties identical to those of PAF. Independent determination of the structures of PAF and APRL showed them to be structurally identical to AGEPC. The commonly accepted name for this substance is platelet-activating factor; however, its actions extend far beyond platelets.

Chemistry and Biosynthesis. PAF is 1-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine.

PAF contains a long-chain alkyl group joined to the glycerol backbone in an ether linkage at position 1 and an acetyl group at position 2. PAF actually represents a family of phospholipids because the alkyl group at position 1 can vary in length from 12-18 carbon atoms. In human neutrophils, PAF consists predominantly of a mixture of the 16- and 18-carbon ethers, but its composition may change when cells are stimulated. The family is expanded by unregulated oxidative production to PAF-like lipids whose mode of action is analogous to PAF (Stafforini et al., 2003).

Analogous to the eicosanoids, PAF is not stored in cells but is synthesized in response to stimulation. The major pathway, termed the remodeling pathway, of PAF generation involves the precursor 1-O-alkyl-2-acyl-glycerophosphocholine, a lipid found in high concentrations in the membranes of many types of cells. The 2-acyl substituents include AA. PAF is synthesized from this substrate in two steps (Figure 33–5). The first involves the action of PLA₂, the initiating enzyme for eicosanoid biosynthesis, with the formation of 1-O-alkyl-2-lyso-glycerophosphocholine (lyso-PAF) and a free fatty acid (usually AA) (Prescott et al., 2000; Honda et al., 2002; Stafforini et al., 2003). Eicosanoid and PAF biosynthesis thus is closely coupled, and deletion of cPLA₂ in mice leads to an almost complete loss of both eicosanoid and PAF generation. The second, rate-limiting step is performed by the acetyl-coenzyme-A-lyso-PAF acetyltransferase. PAF synthesis also can occur de novo; a phosphocholine substituent is transferred to alkyl acetyl glycerol by a distinct lyso-glycerophosphate acetyl-coenzyme-A transferase. This pathway may contribute to physiological levels of PAF for normal cellular functions. The synthesis of PAF may be stimulated during antigen–antibody reactions or by a variety of agents, including chemotactic peptides, thrombin, collagen, and other autacoids; PAF also can stimulate its own formation. Both the phospholipase and acetyltransferase are Ca²⁺-dependent enzymes; thus, PAF synthesis is regulated by the availability of Ca²⁺.

The inactivation of PAF is catalyzed by PAF acetylhydrolyases (PAF-AHs) (McIntyre et al., 2009) (Figure 33–5). This is a family of four related Ca²⁺-independent phospholipases A₂ that show marked specificity for phospholipids with short acyl chains at the sn-2 position. Two PAF-AHs, which are group VII enzymes, are commonly known as plasma PAF-AH (or lipoprotein-associated PLA₂) and liver type II PAF-AH. The remaining two are known as type I and II intracellular PAF-AH. PAF is inactivated by PAF-AH-catalyzed hydrolysis of the acetyl group, generating Lyso-PAF, which is then converted to a 1-O-alkyl-2-acyl-glycerophosphocholine by an acyltransferase.

PAF is synthesized by platelets, neutrophils, monocytes, mast cells, eosinophils, renal mesangial cells, renal medullary cells, and vascular endothelial cells. Depending on cell type, PAF can either remain cell associated or be secreted. For example, PAF is released from monocytes but retained by leukocytes and endothelial cells. In endothelial cells, PAF is displayed on the surface for juxtacrine signaling and stimulates adherent leukocytes (Honda et al., 2002).

In addition to these enzymatic routes, PAF-like molecules can be formed from the oxidative fragmentation of membrane phospholipids (oxPLs) (Stafforini et al., 2003). These compounds are increased in settings of oxidant stress, such as cigarette smoking, and differ structurally from PAF in that they contain a fatty acid at the sn-1 position of glycerol joined through an ester bond and various short-chain acyl groups at the sn-2 position. oxPLs mimic the structure of PAF closely enough to bind to its receptor (see "Mechanism of Action of PAF") and elicit the same responses. Unlike the synthesis of PAF, which is highly controlled, oxPL production is unregulated; degradation by PAF-AH, therefore, is necessary to suppress the toxicity of oxPLs. Levels of plasma PAF-AH are increased in colon cancer, cardiovascular disease, and stroke (Prescott et al., 2002), and polymorphisms have been associated with altered risk of cardiovascular events (Ninio et al., 2004). A common missense mutation in Japanese people is associated disproportionately with more severe asthma.

Mechanism of Action of PAF. Extracellular PAF exerts its actions by stimulating a specific GPCR (Honda et al., 2002). The PAF receptor's strict recognition requirements, including a specific head group and specific atypical sn-2 residue, also are met by oxPLs. The PAF receptor couples to activation of the PLC-IP₃–Ca²⁺ pathway and to inhibition of adenylyl cyclase, via G_q and G_i respectively. Activation of phospholipases A₂, C, and D give rise to second messengers. These include AA-derived PGs, TxA₂, or LTs, which may function as extracellular mediators of the effects of PAF. In addition, p38 MAP kinase is activated downstream of PAF-receptor coupling to G_q, while ERK activation can occur through a number of pathways, including coupling to G_q, G_o, or G_βγ or via transactivation of the epidermal growth factor receptor, leading to NF-κB activation (Stafforini et al., 2003). Thus, the PAF receptor couples to a variety of downstream signaling cascades in a cell-dependent manner.

PAF exerts many of its important pro-inflammatory actions without leaving its cell of origin. For example, PAF is synthesized in a regulated fashion by endothelial cells stimulated by inflammatory mediators. This PAF is presented on the surface of the endothelium, where it activates the PAF receptor on juxtaposed cells, including platelets, polymorphonuclear leukocytes, and monocytes, and acts cooperatively with P selectin to promote adhesion (Prescott et al., 2000). This function of PAF is important for orchestrating the interaction of platelets and circulating inflammatory cells with the inflamed endothelium.

Reproduction and Parturition. A role for PAF in ovulation, implantation, and parturition has been suggested by numerous studies. However, PAF receptor–deficient mice are normal reproductively, indicating that PAF may not be essential for reproduction.

Inflammatory and Allergic Responses. The pro-inflammatory actions of PAF, and its elaboration by endothelial cells, leukocytes, and mast cells under inflammatory conditions, are well characterized. The plasma concentration of PAF is increased in experimental anaphylactic shock, and the administration of PAF reproduces many of its signs and symptoms, suggesting a role for the autacoid in anaphylactic shock. Indeed, studies with PAF receptor–deficient mice confirm the actions of PAF, and PAF-like molecules, as dominant anaphylactic mediators. In addition, mice overexpressing the PAF receptor exhibit bronchial hyperreactivity and increased lethality when treated with endotoxin (Ishii et al., 2002). PAF-receptor knockout mice display milder anaphylactic responses to exogenous antigen challenge, including less cardiac instability, airway constriction, and alveolar edema; they are, however, still susceptible to endotoxic shock. Deletion of the PAF receptor augments the lethality of infection with gram-negative bacteria while improving host defense against gram-positive pneumococcal pneumonia. PAF receptor-overexpressing mice show increased airway responsiveness when challenged, an effect likely mediated through generation of TxA₂ and CysLTs (Ishii and Shimizu, 2000). PAF-AH deficiency is associated with asthma in some populations (Stafforini, 2009).

Despite the broad implications of these observations, the effects of PAF antagonists in the treatment of inflammatory and allergic disorders have been disappointing. Although PAF antagonists reverse the bronchoconstriction of anaphylactic shock and improve survival in animal models, the impact of these agents on animal models of asthma and inflammation is marginal. Similarly, in patients with asthma, PAF antagonists partially inhibit the bronchoconstriction induced by antigen challenge but not by challenges by methacholine, exercise, or inhalation of cold air. These results may reflect the complexity of these pathological conditions and the likelihood that other mediators contribute to the inflammation associated with these disorders.

Cardiovascular System. PAF is a potent vasodilator in most vascular beds; when administered intravenously, it causes hypotension in all species studied. PAF-induced vasodilation is independent of effects on sympathetic innervation, the renin–angiotensin system, or arachidonate metabolism and likely results from a combination of direct and indirect actions. PAF may, alternatively, induce vasoconstriction depending on the concentration, vascular bed, and involvement of platelets or leukocytes. For example, the intracoronary administration of very low concentrations of PAF increases coronary blood flow by a mechanism that involves the release of a platelet-derived vasodilator. Coronary blood flow is decreased at higher doses by the formation of intravascular aggregates of platelets and/or the formation of TxA₂. The pulmonary vasculature also is constricted by PAF, and a similar mechanism is thought to be involved.

Intradermal injection of PAF causes an initial vasoconstriction followed by a typical wheal and flare. PAF increases vascular permeability and edema in the same manner as histamine and bradykinin. The increase in permeability is due to contraction of venular endothelial cells, but PAF is more potent than histamine or bradykinin by three orders of magnitude.

Platelets. The PAF receptor is constitutively expressed on the surface of platelets. PAF potently stimulates platelet aggregation in vitro and in vivo. Although this is accompanied by the release of TxA₂ and the granular contents of the platelet, PAF does not require the presence of TxA₂ or other aggregating agents to produce this effect. The intravenous injection of PAF causes formation of intravascular platelet aggregates and thrombocytopenia.

Because PAF is synthesized by platelets and promotes aggregation, it was proposed as the mediator of COX-inhibitor-resistant, thrombin-induced aggregation. However, no bleeding or thrombotic phenotypes have been associated with deletion of the PAF receptor. In addition, PAF antagonists fail to block thrombin-induced aggregation, even though they prolong bleeding time and prevent thrombus formation in some experimental models. Thus, PAF may contribute to thrombus formation, but it does not function as an independent mediator of platelet aggregation.

Leukocytes. PAF is a potent and common activator of inflammatory cells, providing a point of convergence between the hemostatic and innate immune systems (Zimmerman et al., 2002). PAF stimulates a variety of responses in polymorphonuclear leukocytes (eosinophils, neutrophils, and basophils) when it is generated in both physiological and pathological situations. PAF stimulates PMNs to aggregate, degranulate, and generate free radicals and LTs. Because LTB₄ is more potent in inducing leukocyte aggregation, it may mediate the aggregatory effects of PAF. PAF is a potent chemotactic for eosinophils, neutrophils, and monocytes and promotes PMN-endothelial adhesion contributing, along with other adhesion molecular systems, to leukocyte rolling, tight adhesion, and migration through the endothelial monolayer. PAF also stimulates basophils to release histamine, activates mast cells, and induces cytokine release from monocytes. In addition, PAF promotes aggregation of monocytes and degranulation of eosinophils. When given systemically, PAF causes leukocytopenia, with neutrophils showing the greatest decline. Intradermal injection causes the accumulation of neutrophils and mononuclear cells at the site of injection. Inhaled PAF increases the infiltration of eosinophils into the airways.

Smooth Muscle. PAF generally contracts GI, uterine, and pulmonary smooth muscle. PAF enhances the amplitude of spontaneous uterine contractions; quiescent muscle contracts rapidly in a phasic fashion. These contractions are inhibited by inhibitors of PG synthesis. PAF does not affect tracheal smooth muscle but contracts airway smooth muscle. Most evidence suggests that another autacoid (e.g., LTC₄ or TxA₂) mediates this effect of PAF. When given by aerosol, PAF increases airway resistance as well as the responsiveness to other bronchoconstrictors. PAF also increases mucus secretion and the permeability of pulmonary microvessels; this results in fluid accumulation in the mucosal and submucosal regions of the bronchi and trachea.

Stomach. In addition to contracting the fundus of the stomach, PAF is the most potent known ulcerogen. When given intravenously, it causes hemorrhagic erosions of the gastric mucosa that extend into the submucosa.

Kidney. When infused intrarenally in animals, PAF decreases renal blood flow, glomerular filtration rate, urine volume, and excretion of Na⁺ without changes in systemic hemodynamics (Lopez-Novoa, 1999). These effects are the result of a direct action on the renal circulation. PAF exerts a receptor-mediated biphasic effect on afferent arterioles, dilating them at low concentrations and constricting them at higher concentrations. The vasoconstrictor effect appears to be mediated, at least in part, by COX products, whereas vasodilation is a consequence of the stimulation of NO production by endothelium.

Other. PAF receptor-overexpressing mice spontaneously develop melanocyte tumors. PAF, a potent mediator of angiogenesis, also has been implicated in breast and prostate cancer. PAF-AH deficiency has been associated with small increases in a range of cardiovascular and thrombotic diseases in some human populations (Stafforini, 2009). Administration of PAF-AH reduces injury-induced neointima formation and atherogenesis in genetically prone mice, suggesting a role for PAF in inflammatory cardiovascular disease.