P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`Annu. Rev. Biochem. 2000. 69:419–45
`Copyright c(cid:13) 2000 by Annual Reviews. All rights reserved
`
`PLATELET-ACTIVATING FACTOR
`AND RELATED LIPID MEDIATORS
`
`Stephen M. Prescott1, Guy A. Zimmerman2,
`Diana M. Stafforini1, and Thomas M. McIntyre2
`The Huntsman Cancer Institute1 and the Program in Human Molecular Biology
`and Genetics2, University of Utah, Salt Lake City, Utah 84112;
`e-mail: steve.prescott@hci.utah.edu
`
`Key Words PAF, inflammation, phospholipases
`n Abstract Platelet-activating factor (PAF) is a phospholipid with potent, diverse
`physiological actions, particularly as a mediator of inflammation. The synthesis, trans-
`port, and degradation of PAF are tightly regulated, and the biochemical basis for many
`of these processes has been elucidated in recent years. Many of the actions of PAF can
`be mimicked by structurally related phospholipids that are derived from nonenzymatic
`oxidation, because such compounds can bind to the PAF receptor. This process cir-
`cumvents much of the biochemical control and presumably is regulated primarily by
`the rate of degradation, which is catalyzed by PAF acetylhydrolase. The isolation of
`cDNA clones encoding most of the key proteins involved in regulating PAF has allowed
`substantial recent progress and will facilitate studies to determine the structural basis
`for substrate specificity and the precise role of PAF in physiological events.
`
`CONTENTS
`
`INTRODUCTION : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 419
`PHYSIOLOGIC ACTIONS OF PLATELET-ACTIVATING FACTOR : : : : : : : : : : : 420
`PAF Is a Cell-Associated Signal for Leukocyte Adhesion and Activation
`in Models of Acute Inflammation: Juxtacrine Signaling : : : : : : : : : : : : : : : : : : : 420
`Signaling Roles for PAF in Vivo : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 422
`THE PAF RECEPTOR : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 424
`SYNTHESIS OF PLATELET-ACTIVATING FACTOR
`AND GENERATION OF RELATED COMPOUNDS : : : : : : : : : : : : : : : : : : : : : : 428
`Platelet-Activating Factor : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 428
`Oxidative Generation of Platelet Activating Factor-Like Lipids : : : : : : : : : : : : : : 431
`DEGRADATION OF PAF: THE ACETYLHYDROLASES : : : : : : : : : : : : : : : : : : 433
`The Intracellular Forms : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 435
`The Plasma Form : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 437
`CONCLUSION : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 441
`
`0066-4154/00/0707-0419/$14.00
`
`419
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 1
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`420
`
`PRESCOTT ET AL
`
`INTRODUCTION
`
`Platelet-activating factor (PAF) is the trivial name for a phospholipid, 1-O-alkyl-2-
`acetyl-sn-glycero-3-phosphocholine that has potent, diverse physiological actions.
`Its discovery and the early studies to characterize its actions, synthesis, and degra-
`dation were described in a previous chapter in this series (1) and other reviews
`(2–5). Many lipid mediators are derived from phospholipids (6), but PAF was the
`first one shown to have autocoid, or messenger, functions as an intact phospholipid.
`This signaling results from PAF binding to a specific receptor—not from physico-
`chemical effects on the plasma membrane of the target cell. One of the most in-
`triguing aspects of PAF has been the strict structural requirement for binding to
`its receptor and for recognition as a substrate both by synthetic and degradative
`enzymes. The last point seems to be an eminently practical requirement because, if
`the degradative enzymes were not specific for this unusual structure, there would
`be continuous hydrolysis of structural phospholipids.
`
`PHYSIOLOGIC ACTIONS OF PLATELET-ACTIVATING
`FACTOR
`
`PAF’s primary role seems clearly to be to mediate intercellular interactions; when
`synthesized by cells of a variety of types, it binds to receptors on the plasma
`membranes of other cells, which activates them and changes their phenotypes. In
`addition to intercellular signaling, it has been proposed that PAF has both autocrine
`and intracrine effects, the latter occurring via one or more intracellular receptors.
`Although postulated to exist for many years based on pharmacological studies,
`intracellular receptors for PAF are yet to be rigorously characterized. We focus
`here on intercellular signaling by PAF. Many of its best-characterized signaling
`roles are in the vascular and inflammatory systems (see below), but it also trans-
`mits information between cells in the central nervous system and in endocrine,
`gastrointestinal, and other organs (see review articles above). Most of the follow-
`ing examples are drawn from studies in inflammation, but the principles likely are
`relevant in multiple systems. Several mechanisms regulate the PAF intercellular-
`signaling system. These include tightly controlled synthetic pathways, spatial reg-
`ulation of the display and biologic availability of PAF, cell-specific expression of
`the receptor for PAF, homologous and heterologous desensitization of the receptor,
`and rapid degradation of PAF by extracellular and intracellular acetylhydrolases.
`Each of these regulatory features is considered in more detail elsewhere in the
`chapter. Together, they indicate specialization of function for PAF as an intercel-
`lular signal, and the redundant mechanisms appear to have evolved to precisely
`control its biologic activities. Thus, unregulated or dysregulated signaling by PAF
`can be a mechanism of disease. Disease models in isolated cell systems (7–10) and
`experimental animals (5) and observations in human syndromes (5) also support
`this possibility.
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 2
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`PAF AND RELATED LIPIDS
`
`421
`
`PAF Is a Cell-Associated Signal for Leukocyte Adhesion
`and Activation in Models of Acute Inflammation:
`Juxtacrine Signaling
`In the inflammatory response, PAF has well-characterized actions, mediating cell-
`cell interactions in models of acute and chronic inflammation in virtually all or-
`gans. The cells involved include endothelial cells, leukocytes of several classes, and
`others (5). These studies have shown that PAF can mediate paracrine signaling by
`acting over short distances in solution. In some cases, PAF may also circulate and
`act in an endocrine fashion. The latter mode of action is controlled by plasma PAF
`acetylhydrolase, which limits the half-life of PAF to a few minutes in the blood
`of humans and experimental animals (see below). Experiments in models of acute
`inflammation have also shown, however, that most signaling by PAF may occur be-
`tween closely juxtaposed cells and that it can be recognized by its receptor on target
`cells while associated with the plasma membrane of the signaling cell. Thus, PAF
`can signal in a juxtacrine fashion (Figure 1). These models have also demonstrated
`that PAF acts cooperatively with an adhesion protein, P-selectin, on the endothe-
`lial cell surface and that activation responses triggered by signals delivered via the
`PAF receptor in the target cell further modify the intercellular interaction.
`
`Figure 1 Platelet-activating factor (PAF) signals the priming and activation of leukocytes
`at the surfaces of inflamed human endothelial cells (a juxtacrine system for spatial control of
`inflammation). PAF and an adhesion protein, P-selectin, are coordinately displayed on the
`plasma membranes of stimulated human endothelial cells. P-selectin tethers the leukocyte
`to the endothelial cell, which allows the PAF from the endothelial cell to bind to its receptor
`on the polymorphonuclear leukocyte. This constitutes a form of juxtacrine signaling, which
`may be a general way to spatially restrict the actions of a potent, pleiotropic mediator such
`as PAF. (cid:12)2 integrins are not shown on the polymorphonuclear leukocyte (left side of the
`figure), for convenience. This figure is reproduced from previously published work (10a),
`with permission of the publisher.
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 3
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`422
`
`PRESCOTT ET AL
`
`PAF had been shown in early studies to be a fluid-phase mediator that could
`activate platelets and various leukocytes in suspension (1). Subsequently, it was
`demonstrated that PAF is synthesized in regulated fashion by endothelial cells
`stimulated with thrombin or other inflammatory mediators and that this PAF ac-
`tivated polymorphonuclear leukocytes (PMNs) (11–13), which are key effector
`cells in the acute inflammatory response. This was the first observation of synthe-
`sis of a signaling factor for PMNs by inflamed endothelial cells, and it established
`a molecular mechanism for local activation of leukocytes at the endothelial sur-
`face rather than requiring the diffusion of chemotactic factors into the blood. This
`local mechanism of leukocyte signaling had been predicted by classic in vivo
`studies but never proven. The adhesion of PMNs in response to PAF is mediated
`by activation-dependent alterations in the affinity and avidity of (cid:12)2 integrins on
`their surface (14). PAF also induces other responses of PMNs that are critical in
`acute inflammation, including cell polarization, enhanced motility, priming for en-
`hanced granular enzyme release, and redistribution of surface ligands (Figure 1).
`In addition, PAF activates other leukocytes involved in inflammation, such as
`monocytes, where PAF signals NF-(cid:20)B translocation to the nucleus and alterations
`in gene expression (15) and other functional responses.
`A key additional observation was that PAF synthesized by stimulated endothe-
`lial cells is not released into solution but that almost all of it remains cell-associated
`even in the presence of albumin, an acceptor molecule to which it binds avidly
`(13, 16, 17). Furthermore, a significant fraction of PAF that is synthesized by stim-
`ulated endothelial cells is translocated to the outer surface of the plasma membrane,
`where it is available for binding to its receptor on target PMNs (18). The specific
`mechanism of translocation to the endothelial surface and whether PAF is local-
`ized in hydrophobic patches or is noncovalently linked to a presenting protein are
`currently unknown. A variety of experimental strategies were used to demonstrate
`that PAF that is endogenously synthesized by endothelial cells and is retained at
`their surfaces ligates the PAF receptor on target PMNs. These strategies included
`receptor blockade, receptor desensitization, and degradation of PAF in the endothe-
`lial plasma membrane with extracellular preparations of PAF acetylhydrolase (18).
`These experiments demonstrated that PAF displayed by inflamed endothelial cells
`has the requisite features of a juxtacrine signaling molecule (19). Subsequent ex-
`periments have confirmed that PAF can operate in this fashion when displayed by
`stimulated endothelial cells both in static systems and when the target PMNs are
`subjected to flow to model in vivo conditions (20–23; reviewed in 24). Recently,
`strategies with exogenous recombinant PAF acetylhydrolase (25) or PAF receptor
`antagonists (26) demonstrated that PAF is a juxtacrine signaling molecule at the
`surfaces of activated human platelets, which—like endothelial cells—are critical
`in cell-cell interactions in inflammatory and thrombotic responses (27, 28). These
`findings suggest that this mode of action, presentation of PAF on the cell surface
`by the cell originating the signal with transfer to an adjacent cell that has a recep-
`tor, may be a general mechanism. If so, it has profound implications for how this
`mediator and other small signaling molecules function in many tissues. In essence,
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 4
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`PAF AND RELATED LIPIDS
`
`423
`
`it would make the concentration, as determined in biological fluids such as blood,
`irrelevant because the crucial determinant would be its relative abundance in a
`pseudo-two-dimensional space. This could have attractive implications for signal-
`ing events such as occur in the neuronal synapse and other environments where
`spatial limitation of the signal would be important.
`
`Signaling Roles for PAF in Vivo
`Many studies involving exogenous administration of PAF or its endogenous gener-
`ation in experimental animals indicate that it has signaling roles in vivo (reviewed
`in 5). Again, many of these involve models of inflammation or inflammatory injury.
`Recent studies with genetically manipulated mice (29, 30) further document signal-
`ing roles of PAF in vivo. In the first, the guinea pig PAF receptor was overexpressed
`in mice by Shimizu and colleagues (29), who originally cloned the guinea pig re-
`ceptor (31). The transgenic animals had increased mortality when challenged with
`endotoxin, developed melanocytic tumors of the skin, and had increased bronchial
`hyperreactivity in response to inhaled methacholine. Increased bronchial reactivity
`and susceptibility to endotoxin were consistent with previous studies in wild-type
`animals, where there is considerable evidence that dysregulated signaling by PAF
`or PAF-like lipids occurs in endotoxemia and models of asthma (5). Melanocytic
`tumors were, however, unexpected although there is evidence that PAF has neo-
`plastic effects, and it is known to mediate skin inflammation. In a converse strategy,
`the same group created mice with targeted deletions in the PAF receptor (30). These
`PAF receptor knockout animals were developmentally normal and reproduced ef-
`fectively, indicating that some of the actions in which PAF had been implicated pre-
`viously are not essential actions. Future studies with these mice are important be-
`cause more sensitive assays might detect subtle alterations in reproductive capacity,
`neurological function, and other normal physiological events in which PAF might
`play a role. In the reported studies, the knockout animals were tested in a model
`of inflammation and had much milder anaphylactic responses to exogenous anti-
`gen challenge than did wild-type animals, including less cardiovascular instability,
`airway constriction, and alveolar edema. These results were again consistent with
`findings in earlier animal models, involving receptor blockade or other manipula-
`tions of the PAF system, and were also consistent with the respiratory physiology
`of the transgenic mice overexpressing the PAF receptor. In contrast, the knockout
`animals remained susceptible to endotoxin, with vascular and cytokine responses
`equivalent to those in the paired animals with intact PAF receptors. This suggested
`that PAF is not required for endotoxic shock, although it is a modulating signal
`(30). A potentially important variable is that the genetic background of mice used in
`this study was different from that of the transgenic mice that overexpressed to PAF
`receptor.
`In a third study, Shimizu and colleagues examined inflammatory acute lung
`injury caused by acid aspiration in knockout mice and in transgenic mice overex-
`pressing the PAF receptor. The injury was reduced in animals deficient in the PAF
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 5
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`424
`
`PRESCOTT ET AL
`
`receptor and increased in the overexpressing mice when assessed by histologic and
`physiologic indices and mortality (10). An unexpected finding was that local accu-
`mulation or sequestration of leukocytes in pulmonary vessels of acid-injured mice
`was not different from that of the wild type in either the knockout or transgenic
`animals. However, this is consistent with evidence that there are mechanisms of
`adhesion and accumulation of leukocytes that are independent of activation of (cid:12)2
`integrins in the mouse pulmonary system. The fact that injury was reduced sug-
`gests that signaling by the PAF receptor to priming pathways that enhance granular
`enzyme release or oxygen radical generation (5), both of which are known to cause
`lung injury, is critical in this model.
`Studies in humans also support signaling roles for PAF in vivo. Many reports of
`individual patients and small groups of subjects indicate that dysregulated signal-
`ing by PAF may be a mechanism of disease (5, 32). As examples, developmental
`deficiency of PAF acetylhydrolase may lead to unregulated proinflammatory sig-
`naling by PAF in infants with neonatal necrotizing enterocolitis and in laboratory
`animals with an experimental form of this condition (32–34), and dysregulated sig-
`naling by PAF contributes to the severity of asthma in patients genetically deficient
`in PAF acetylhydrolase (35).
`
`THE PAF RECEPTOR
`
`The intercellular actions of PAF are mediated through a G-protein–linked receptor
`that is expressed on the surface of a variety of cell types (31, 36–38). The human
`gene codes for 342 amino acids, and the sequence is >80% identical to the guinea
`pig protein. After the isolation of the receptor cDNA, the signaling pathway was
`defined in transfection experiments, which complemented previous pharmacolog-
`ical studies (39). Ali et al (40) showed that the signal from the PAF receptor caused
`phosphatidylinositol turnover and raised the intracellular calcium. Both responses
`were inhibited by GTP analogs, but had different responses to pertussis toxin. This
`suggested that there were two different G proteins downstream of the PAF receptor.
`Honda et al (41) used stably transfected cells to assess the role of mitogen-activated
`protein (MAP) kinases and described activation of the 42- and 44-kDa forms of
`ERK by tyrosine phosphorylation when the cells were exposed to PAF. They, too,
`observed increased phosphatidylinositol turnover and found G-protein pathways
`that were sensitive and insensitive to pertussis toxin. They were unable to detect
`activation of ras, which indicated that MAP kinase activation occurred through a
`different pathway (41).
`The structure-function relationships of the PAF receptor have been defined by
`site-directed mutagenesis and transfection (Figure 2). One goal was to identify
`the PAF-binding site, and, based on mutagenesis studies, Ishii et al proposed that
`histidine residues 188, 248, and 249 form a binding pocket (42) and systematically
`mutated all of the polar amino acids in the transmembrane domains of the receptor
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 6
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`PAF AND RELATED LIPIDS
`
`425
`
`permissionofthepublisher.
`ligandbinding,intracellularsignaling,anddesensitization.Thisfigureisreproducedfrompreviouslypublishedwork(5),with
`cDNAsencodingwild-typeandmutantreceptorsintotargetcellshavedefinedtheregionsofthereceptorresponsiblefor
`Figure2Thestructure-activityrelationshipsoftheplatelet-activatingfactor(PAF)receptor.Studiesusingtransfectionof
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 7
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`426
`
`PRESCOTT ET AL
`
`(Figure 2) and found that mutations in domains II, III, and VII resulted in higher
`affinities for PAF than the wild-type receptor. Conversely, mutants in transmem-
`brane segments V, VI, and VII had lower affinities (42). The binding affinities
`correlated well with functional responses. A mutation in the third transmembrane
`domain was constitutively active and had lost some of the substrate specificity; the
`mutant receptor responded well to lyso-PAF, which gives little or no response with
`the wild-type receptor. Parent et al (43) examined the transmembrane domains
`and found two adjacent phenylalanines (Figure 2) to be important for structural
`stability of the receptor but, surprisingly, they had no substantial effect on lig-
`and binding (43). They also explored the role of several cysteine residues and
`showed that there likely is a disulfide bond between the conserved residues at
`positions 90 and 163 (44). The cysteine at position 95 was required for optimal
`function, and they concluded that it, too, might be bound to another cysteine.
`In addition to the mutation (N100A) described by Ishii et al, which results in
`constitutive activation of the PAF receptor, Parent et al showed that mutation of
`alanine 230 to glutamic acid also yielded constitutive activity and an increased
`affinity for binding of PAF (45). Mutation of the adjacent residue led to a marked
`decrease in activity. Thus, the central portion of the receptor appears to func-
`tion in ligand binding. This type of binding site is consistent with findings for
`other G-protein-coupled receptors, in which the ligand-binding site is defined in
`three dimensions by amino acids from several of the transmembrane-spanning
`regions.
`Another series of studies defined the portions of the receptor responsible for
`transducing PAF binding into intracellular signals. The third intracellular loop has
`been implicated by studies from multiple laboratories. Carlson et al (46) showed
`that constructs with the sequence from this region could block the response from
`other receptors (this essentially was a “dominant negative” strategy to saturate the
`downstream signaling pathways). They also showed that chimeric receptors that
`included or lacked this region could shift the signaling response. From these ex-
`periments, it seems clear that the third intracellular loop of the receptor is critically
`important for initiating phosphatidylinositol turnover.
`As with other G-protein–coupled receptors, the PAF receptor is desensitized
`after its initial activation. Ali et al (40) found that desensitization of the PAF recep-
`tor was accompanied by phosphorylation, which was partially blocked by protein
`kinase C (PKC) inhibitors. PKC is activated by PAF, so their finding suggested
`a negative feedback mechanism to desensitize the receptor (40). This was ex-
`plored by Takano et al (47), who created a PAF receptor that lacked the C-terminal
`intracellular region and another in which several serine residues and one threo-
`nine residue were replaced with alanine. Both the tail region and these putative
`phosphorylation sites were required for desensitization. Unexpectedly, however,
`pretreatment either with phorbol esters or PKC inhibitors had no effect, which
`indicated that PKC was not the responsible kinase. They also found that a peptide
`representing the carboxyl-terminal 18 amino acids of the receptor was an excellent
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 8
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`PAF AND RELATED LIPIDS
`
`427
`
`substrate for the G-protein–coupled receptor kinase-1 (47). Taken together, these
`experiments showed that the carboxyl terminus of the PAF receptor is the target for
`phosphorylation and that this is required to desensitize the receptor. The respon-
`sible kinase is likely to be a member of the G-protein–coupled receptor kinase-1
`family.
`At a cellular level there may be additional mechanisms for desensitization of
`responses to PAF—a phenomenon known as heterologous desensitization. In the
`down-regulation described above (homologous desensitization), receptors are de-
`sensitized only by their own ligand (or family members). In contrast, in heterolo-
`gous desensitization, receptors of one type can be down-regulated after activation
`of a different type of receptor. From previous work, it was known that the PAF
`receptor undergoes heterologous desensitization when some peptide receptors are
`activated, and other work showed that this results from PKC-mediated phosphory-
`lation at the carboxyl terminus (48). The converse can also be true—signals from
`the PAF receptor can down-regulate peptide receptors by changes in downstream
`enzymes (49). In summary, the PAF receptor can be desensitized in multiple way;
`homologous desensitization of PAF uses two mechanisms—phosphorylation of
`the receptor (probably a G-protein–coupled receptor kinase-1) and inactivation
`of the key downstream effector PLC-(cid:12)3 by PKC. Heterologous desensitization
`of the PAF receptor uses a third mechanism—phosphorylation of the receptor by
`PKC.
`The structure of the human gene encoding the PAF receptor has been deter-
`mined, and it does not contain introns (50, 51). The gene is on chromosome
`1p35-p34.4 (50, 51). In cells that normally express the PAF receptor, this gene
`seems to be present constitutively, but it can be transcriptionally regulated by
`inflammatory agonists including PAF itself (52–55). There are two different tran-
`scripts, which are found in different tissues. The first transcript described is found
`only in peripheral leukocytes and is the one that increases in response to PAF.
`It has consensus sequences for NF-(cid:20)B, SP-1, and INR, and the response to PAF
`and phorbol esters requires the NF-(cid:20)B consensus binding sites (53). Transcrip-
`tion of the second transcript is induced by phorbol esters, but not PAF, and this
`response requires a functional AP-1 site (53). The second transcript also is reg-
`ulated by retinoic acid and thyroid hormone (56). The cis-acting elements re-
`sponsible for this effect were mapped to three hexamer repeats located between
`−67 to −44 of the transcription start site. Thus, the regulation of expression of
`0
`sequence of transcript 1 re-
`the two transcripts seems to be quite specific—the 5
`sponds to inflammatory stimuli, whereas transcript 2 is regulated by differentiation
`signals.
`It seems likely that the major control of expression of the PAF receptor is at
`the transcriptional level, but one group observed that the PAF receptor mRNA
`levels in human monocytes fell in response to PKC and that the t1=2 for the mRNA
`was markedly decreased (52). They concluded that the stability of the mRNA was
`another potential regulatory mechanism, and this warrants further study.
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 9
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`428
`
`PRESCOTT ET AL
`
`SYNTHESIS OF PLATELET-ACTIVATING FACTOR
`AND GENERATION OF RELATED COMPOUNDS
`
`Platelet-Activating Factor
`PAF may be synthesized through either of two enzymatic pathways, one defined as
`the remodeling pathway that substitutes an acetyl residue for the long-chain fatty
`acyl residue (Figure 3) of cellular phospholipids. A second de novo pathway to
`form PAF parallels phospholipid synthesis, in which a phosphocholine function is
`transferred to alkyl acetyl glycerol (5). Recent evidence supports the former path-
`way as contributing the bulk of the PAF synthesized in an inflammatory setting.
`PAF, like the eicosanoids, is not stored in a preformed state, but rather is rapidly
`synthesized by inflammatory cells in response to cell-specific stimuli. Stimulated
`PAF synthesis through the remodeling pathway begins with the activation of a
`phospholipase that hydrolyzes phosphatidylcholines to radyl (an inclusive term
`that does not distinguish what type of bond is at the sn-1 position of phospho-
`lipids) lysophosphatidylcholines. This enzyme, cPLA2, is regulated by Ca2C
`and
`by phosphorylation by MAP kinases (57). The stimulation by Ca2C
`occurs at sub-
`micromolar concentrations, and so its activity will be modulated by the intracellular
`Ca2C
`changes found in stimulated cells. This enzyme is unusual in that it displays
`a marked specificity for an sn-2 arachidonoyl residue in contrast to the broadly
`permissive nature of most of the members of the phospholipase A2 super family.
`This enzyme is responsible for nearly all of the arachidonate release necessary for
`eicosanoid synthesis, but it also is an essential first step in the synthesis of PAF.
`Both points were suggested by antisense inhibition of the enzyme in two types
`of cell lines (58, 59) and then firmly established by the nearly complete loss of
`prostanoid (60) and PAF synthesis by the elicited peritoneal macrophages of two
`independently established cPLA2 knockout animals (61, 62). This enzyme, then,
`is essential for arachidonate and radyl glycerophosphocholine release in inflam-
`matory settings. It does not distinguish what structure is present at the sn-1 posi-
`tion (63) and therefore attacks alkyl phosphatidylcholines in addition to the more
`abundant diacyl species. This implies that the initial lysophospholipid products re-
`flect the sn-1 alkyl and acyl arachidonoyl phosphatidylcholine composition of the
`cell.
`The cellular contents of 1-O-alkyl-arachidonoyl-sn-glycero-3-phosphocholine
`and 1-fatty acyl-arachidonoyl-l-sn-glycero-3-phosphocholine vary over a wide
`range. Alkyl phospholipids are trace components of the phosphatidylcholine pool
`of most cells, but, in cells that synthesize PAF such as endothelial cells (64) and
`neutrophils (65), the alkyl phosphatidylcholine content ranges from 10% to 40%,
`respectively, of the total cellular choline phosphoglyceride pool. In leukocytes and
`monocytes, the alkyl phosphatidylcholines are enriched for arachidonate (66, 67),
`where (cid:24)60% of the cellular arachidonate is found in this species. Accordingly,
`this subclass of phosphatidylcholine accounts for 60% of the arachidonate released
`from stimulated cells (66, 67). As would be predicted from the role of cPLA2 in
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 10
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`PAF AND RELATED LIPIDS
`
`429
`
`andsoarepotentligandsandagonistsofthePAFreceptor.
`fragmentsandmayoxidizethesn-2residue.SomeofthemanyphospholipidoxidationproductscontainthesamefeaturesofPAF
`ashortsn-2residue,andaphosphocholineheadgroup.Uncontrolledoxidation(right)ofalkylpolyunsaturatedphosphatidylcholines
`arachidonoylresidue.ThePAFreceptorrecognizesthreefeaturesofthisphospholipidhormone(shadedarea):ansn-1etherbond,
`synthesisofPAF(left)isacarefullycontrolledprocessinwhichanacetylresidueissubstitutedforapolyunsaturatedsn-2
`Figure3Enzymaticsynthesisofplatelet-activatingfactor(PAF)andnonenzymaticgenerationofPAF-likephospholipids.The
`
`Annu. Rev. Biochem. 2000.69:419-445. Downloaded from www.annualreviews.org
`
` by John Jones on 10/12/17. For personal use only.
`
`AKER EXHIBIT 2003 Page 11
`
`

`

`P1: FPP
`July 25, 2000
`
`12:47
`
`AR102
`
`CHAP15
`
`430
`
`PRESCOTT ET AL
`
`the release of esterified arachidonate in stimulated cells, leukocytes depleted of
`arachidonate by culture in arachidonate-deficient medium showed an 85% de-
`crease in PAF synthesis. All of this shows a strong correlation between PAF and
`eicosanoid synthesis as cells specialized for PAF synthesis store the precursors
`for these two types of lipid messengers together. It also suggests that one way
`to control the formation of PAF, compared with its much less active diacyl ho-
`molog, is by maintaining a high ratio of arachidonate in the sn-2 position of alkyl
`phosphatidylcholines.
`The second and final step in the synthesis of PAF through the remodeling path-
`way is performed by the acetyl-CoA-lysoPAF acetyltransferase. Little is known
`about this activity because this unstable enzyme remains unpurified and uncharac-
`terized, although it is stimulated in a Ca2C
`and phosphorylation-dependent fashion
`(68–70). In leukocytes, the relevant kinase appears to be the p38 MAP kinase (71).
`This acetyltransferase activity possesses little specificity for the sn-1 bond, and
`it will form either PAF from 1-alkyl-sn-glycero-3-phosphocholine (lysoPAF) or
`lysophosphocholine in in vitro assays (72). This lack of substrate selection for the
`sn-1 bond or the presence of two activities is reflected by the relative abundance
`of the acetylated products made by activated cells. For example, the abundance of
`alkyl phosphatidylcholine in endothelial cells is 10% (64), and the relative amount
`of PAF and acylPAF made by these cells is also arou