Molecular Targets for Novel Antiasthma Agents |
 |
INTRODUCTION |
|---|
TOP
INTRODUCTION
CONCLUSION
REFERENCES
|
|---|
Four decades have passed since Sutherland and Rall (1) reported the presence of cyclic nucleotide hydrolyzing activity in extracts of a variety of tissues. This enzymatic activity terminates the biologic actions of cyclic AMP and cyclic GMP, ubiquitous second messengers that regulate countless biologic processes. We now recognize that a large and diverse family of enzymes, called cyclic nucleotide phosphodiesterases (PDEs), is responsible for the cyclic nucleotide hydrolyzing activity described by Sutherland and Rall.
Interest in PDEs as drug targets has grown steadily since they were first described. Several key discoveries fueled this trend. The first discovery was made in the early 1970s when various investigators used a number of standard chromatographic techniques to isolate PDE activities from tissue homogenates (2, 3). These studies revealed that PDE activities from crude tissue extracts eluted from anion-exchange columns in several peaks. Each peak of activity contained at least one kinetically distinct PDE, and, importantly, the relative amounts of the different PDEs varied among different tissues. These observations led to the proposal that PDEs constitute an enzyme family and further suggested these individual members of this family have distinct tissue distributions (4). An obvious extension of these proposals was that selectively targeting different PDE activities with inhibitors could lead to the development of pharmacologic agents possessing a degree of organ or tissue specificity. Notwithstanding the attractiveness of this concept, the likelihood of identifying selective inhibitors represented a major unknown.
By the late 1970s and early 1980s it became clear that kinetically distinct PDEs could indeed be inhibited selectively by a variety of small organic molecules (5). These findings triggered the establishment of a number of research efforts focused on the de novo synthesis of more potent and selective PDE inhibitors. More recently, the application of molecular biology to the identification of new PDEs has presented both unprecedented opportunity and bewildering complexity to drug discovery efforts, in that it is now recognized that PDEs represent a diverse superfamily composed of at least seven genetically distinct isozyme families (9).
Theophylline, which has been used in the treatment of
asthma and other pulmonary diseases for more than 70 yr,
nonselectively inhibits all PDE isozymes in vitro. Although
PDE inhibition might contribute to the therapeutic activity of
theophylline, the relative importance of this mechanism in
comparison with its myriad of other actions is debatable (12-
14). On the other hand, a view that generates less controversy
is that the side effects produced by elevated plasma concentrations of theophylline stem from its tendency to inhibit indiscriminately all PDEs in all tissues of the body (14, 15). Realistically, the true therapeutic potential of PDE inhibition as a
molecular mechanism for antiasthmatic drugs has not been
evaluated with theophylline, as the side effects produced by
nonselective inhibition of all PDEs in all tissues limits the degree of enzyme inhibition achievable in target tissues (i.e., immunocompetent cells, airway smooth muscle). Thus, the concept emerged that selective inhibition of the appropriate PDE isozyme
presumably one that predominates in immune and
inflammatory cellswould produce substantial therapeutic
activity without the side effects that accompany nonselective
PDE inhibition (16).
The purpose of this review is to describe the latest developments in PDE research and to present the scientific rationale that underpins the proposed utility of isozyme selective PDE inhibitors in the treatment of asthma.
| |
OVERVIEW OF PHOSPHODIESTERASES |
|---|
Characteristics of the seven families of PDE isozymes are reviewed comprehensively elsewhere (10, 11, 17) and are summarized in Table 1. The major families are designated by Arabic numerals. Most of the families include more than one gene product as well as multiple splice variants. The GenBank nomenclature for PDEs includes the species and gene family designation (e.g., HSPDE4 for Homo sapien PDE isozyme 4), followed by the gene product or "subtype" (e.g., HSPDE4D) and the splice variant (e.g., HSPDE4D3) (18). The families are also named according to their kinetic characteristics or the identity of endogenous activators (e.g., calmodulin) or inhibitors (e.g., cyclic GMP).
|
Structural and Functional Domains
Phosphodiesterases contain three functional domains. As illustrated in Figure 1, these domains include a conserved catalytic core, a regulatory N-terminus and the C-terminus (11, 19, 20). Considerable sequence similarity (50% or greater identity at the amino acid level) exists among the catalytic regions of PDEs from different families, whereas sequences for the N-termini and C-termini are highly heterologous. The N-terminus serves a regulatory role in several of the PDE families. For example, this region contains a calmodulin-binding domain in PDE1, cyclic GMP-binding sites in PDE2, phosphorylation sites for various protein kinases in PDE1, PDE3, PDE4, and PDE5, and a transducin-binding domain in PDE6. The N-terminus of certain PDEs also contains a membrane targeting domain, which is important in determining cellular and functional compartmentation. Specific functional roles of the C-terminal region are less clear, although a recent report suggests that this domain is important for dimerization of PDE4D1 (21). The three domains are connected by putative hinge regions. Flexibility in these areas may allow the N- or C-terminal domains to fold onto the catalytic region, thus modulating access of the substrate to the catalytic site (11). This affords a mechanism whereby allosteric regulators can increase or decrease enzyme activity by altering the tertiary structure to relieve or impose steric inhibition.
|
Kinetic Characteristics and Allosteric Regulation
All PDEs inactivate cyclic nucleotides by hydrolytically cleaving the 3'-phosphoester bond to form the corresponding inactive 5'-nucleotide monophosphate products. From a functional standpoint the two major differentiating factors among the PDEs are their relative affinities for cyclic AMP and cyclic GMP, and their sensitivities to endogenous and exogenous regulators (Table 1). Regarding substrate affinities, PDE4 and PDE7 are highly selective for cyclic AMP. Although PDE3 hydrolyzes cyclic AMP and cyclic GMP with equal Kms, its Vmax for cyclic AMP is fivefold greater than for cyclic GMP (22). Functionally, then, PDE3 favors cyclic AMP. Cyclic GMP is the preferred substrate for PDE5 and PDE6, whereas PDE1 and PDE2 hydrolyze either cyclic nucleotide. As presented in Figure 2, the PDEs responsible for hydrolyzing cyclic AMP in intact tissues are likely to be different from those responsible for hydrolyzing cyclic GMP.
|
Adding further complexity to the functional role taken by various PDEs in intact tissues is the fact that several are subject to short-term allosteric regulation by endogenous activators or inhibitors (10, 17, 19, 23). For example, PDE1 is allosterically activated by Ca2+/calmodulin. Consequently, PDE1 is activated in intact cells by agents that stimulate Ca2+ mobilization and, consequently, elevate cytosolic Ca2+/calmodulin content (24, 25). Increasing cyclic GMP content in intact cells allosterically activates PDE2, triggering a reciprocal drop in cyclic AMP content (26, 27). On the other hand, cyclic GMP competitively inhibits cyclic AMP hydrolysis by PDE3 (22, 28). This appears to be the mechanism whereby agents that elevate cyclic GMP content in vascular smooth muscle and platelets produce a parallel increase in cyclic AMP content (29, 30). This mechanism may also influence the responsiveness of human airway smooth muscle to inhibitory nonadrenergic-noncholinergic (iNANC) stimulation in the presence of isozyme-selective PDE inhibitors (31).
Protein phosphorylation is another short-term regulatory mechanism for PDEs. Phosphodiesterase 1A1 and PDE1A2 are phosphorylated by protein kinase A (PKA), whereas PDE1B is phosphorylated by calmodulin kinase II (32, 33). In all cases phosphorylation reduces the affinity of the enzyme for Ca2+/calmodulin, which, in effect, decreases the Ca2+ sensitivity of the enzyme. Both subtypes of PDE3 are substrates for PKA; PDE3B is also a substrate for insulin-sensitive protein kinase(s) (17, 23). Phosphorylation of this enzyme by either kinase enhances its activity. Only selected subtypes of PDE4D, those that contain a PKA phosphorylation consensus site within the N-terminal domain, are activated through the PKA pathway (34). This allows differential regulation within a single PDE subtype. Finally, PDE5 is phosphorylated by both PKA and protein kinase G (35, 36). The functional consequences of this phosphorylation are unclear, but it may serve to increase enzyme activity (37).
Inhibitors
Isozyme-selective inhibitors are available for most PDE families. A sampling of these appears in Table 2. More detailed descriptions of the nature of these agents can be found elsewhere (10, 38). In general, these compounds are at least 30-fold selective for the PDE against which they are targeted. Most are substrate-site-directed competitive inhibitors, but a few act at allosteric sites (10, 38). Selective inhibitors of PDE7 have yet to be identified. In fact, this enzyme is resistant to all standard PDE inhibitors, including nonselective compounds such as 3-isobutyl-1-methylxanthine (IBMX) (39, 40).
|
| |
MOLECULAR BIOLOGY |
|---|
Each of the seven families of PDE isozymes is populated by at least one, and as many as four, distinct gene products (Table 1). Furthermore, alternative mRNA processing permits the production of multiple splice variants for each gene. The result of this diversity is a staggering array of enzymes with distinct kinetic characteristics, regulatory properties, and subcellular distributions (10, 11, 17). An in-depth description of the molecular biology of all enzymes within the superfamily of PDEs is beyond the scope of this review. Nonetheless, further insights into the diversity of PDEs can be gained by examining the molecular biology of one of the families, PDE4, in more detail. This enzyme family is particularly pertinent because, as detailed later, it is an important molecular target for novel antiasthmatic drugs (16, 41, 42).
Occurrence and Nature of Multiple PDE4 Gene Products
Four human PDE4 subtypes have been cloned and expressed (43) (Table 3). Sequence comparison indicates that members of the PDE4 family contain two upstream conserved regions (UCR1 and UCR2) as well as the conserved catalytic core (45) (Figure 3). Compared with PDE4A, the other subtypes have amino acid identities of 70 to 74% across the entire sequence and identities of 80 to 84% within the catalytic region (44, 48). Superficially, the PDE4 subtypes have similar kinetic characteristics (cyclic AMP Km = 1.5 to 18 µM and cyclic GMP Km > 1,000 µM) and each is inhibited by rolipram and other archetypical PDE4 inhibitors (44).
|
|
Several studies have indicated that PDE4 subtypes have distinct tissue and cellular distributions. For example, message for PDE4C is abundant in neuronal tissue but is absent from immune and inflammatory cells (49, 50). Message for PDE4A appears to be distributed ubiquitously (45, 47, 50, 51), although in monocytes and perhaps in other cells the expression of this subtype is very low (52). PDE4B message is expressed in heart, brain, skeletal muscle, and lung, but not in placenta, liver, kidney, or pancreas (44).
Significance of PDE4 Splice Variants: Critical Role for the N-Terminal Domain
Additional heterogeneity in the human PDE4 family is produced by alternate mRNA splicing of the PDE4A, PDE4B, and PDE4D loci (18, 20, 56, 57). These alternative splice points generally occur at one of two consensus regions within the 5'-end of the transcripts (20), although an insert occurring in the catalytic domain of PDE4A has also been reported (56). The N-terminal region of PDE4 is important for several functional attributes of the enzyme, including cellular distribution (57), subcellular localization (57, 58), catalytic activity (57, 58), inhibitor binding (46, 48), and regulation by phosphorylation (59, 60). With regard to organelle targeting, removal of the first 67 nucleotides from RD1, a cDNA encoding a rat homolog of human PDE4A, yields an expressed protein that is primarily cytosolic, whereas the full-length protein is membrane-bound (61, 62). Other evidence indicates a differential subcellular distribution of two rat PDE4B splice variants (52). The differential targeting of these splice variants may involve a membrane anchoring domain on the N-terminus of full-length proteins that is lacking in truncated versions (58). Another potential mechanism for compartmentalization is suggested by the presence of SH3 domain-binding sites on PDE4A (63). SH3 domains are found on a number of tyrosyl protein kinases as well as on cytoskeletal proteins (64). Thus, the complementary SH3 domain-binding site on PDE4A may serve to recruit the enzyme into multimeric signaling complexes with specific functional roles (58).
In addition to determining the subcellular distribution of PDE4, the N-terminal domain has a role in modulating catalytic activity. Kinetic analyses of several N-terminal splice variants of rat PDE4A reveal that even though all catalytically active species have the same Km for cyclic AMP, they differ markedly with respect to their Vmax (61, 65). Specifically, the Vmax is the greatest in the shortest form of the enzyme examined, one that is missing nearly half of its N-terminus. Kinetic analyses of full length and truncated forms of human recombinant PDE4A4 reveal a similar pattern (48). The influence of the N-terminus on catalytic activity is not limited to the PDE4A subtype as phosphorylation of Ser54 on PDE4D3 by PKA increases the enzyme's Vmax but has no effect on its Km (60, 66). Two other PDE4D splice variants, each of which is approximately 200 amino acids shorter than PDE4D3 at the N-terminus, are neither substrates for PKA nor activated by PKA (67). These observations led Conti and coworkers (11) to propose that the N-terminus of PDE4D3 contains a regulatory domain that inhibits catalytic activity, perhaps by sterically limiting access of the substrate to the catalytic site. Presumably, phosphorylation of Ser54 induces a conformational change that releases the steric inhibition. Thus, variants of PDE4D and, by analogy, PDE4A that are truncated at the N-terminus might have greater Vmax because they lack the putative inhibitory domain. Nonetheless, the biologic implication of this behavior is that cells could regulate their PDE4A and PDE4D catalytic capacities through phosphorylation or by changing the expression levels of specific splice variants. Indeed, activation of PDE4D3 via phosphorylation is an important mechanism whereby cyclic AMP content is regulated in U937 cells (59), a human monocytic cell line, and FRTL-5 cells (66), a rat thyroid cell line.
The N-terminus of PDE4 also influences inhibitor binding. Kinetic and inhibitor binding analyses conducted on full-length PDE4A4 suggest that two conformational states of the enzyme coexist (46, 48, 68). Rolipram binds to the catalytic sites of both conformers, but with substantially different affinities (Kd ~ 2 nM versus Kd ~ 100 nM). Studies on various truncated versions of PDE4A4 prepared by genetic engineering suggest that amino acid residues 1 through 332 are necessary for the enzyme to assume the conformational state that binds rolipram with high affinity (48). Although catalytic activity is retained or even enhanced with this N-terminal deletion, the potency of rolipram is reduced 10- to 30-fold (48). In contrast, this deletion has no effect on the affinity (Km) of cyclic AMP or the Kd of a structurally distinct PDE4-selective inhibitor, RP 73401 (48). Thus, the N-terminus of PDE4A influences the potency of some, but not all, PDE4 inhibitors. A somewhat analogous situation is seen with PDE4D. As described previously, phosphorylation of Ser54 by PKA increases the activity of PDE4D3 (60, 66). Intriguingly, this phosphorylation also causes a dramatic increase in the potency of rolipram and two other PDE4-selective inhibitors (60, 66). Hence, as with PDE4A, the N-terminal domain of PDE4D is important in regulating the sensitivity of the enzyme to inhibitors. This has important implications with regard to drug targeting. Specifically, since the expression pattern and/or the phosphorylation state of PDE4 N-terminal splice variants differs among tissues (57), the responsiveness of different tissues to inhibitors may vary as well.
| |
ROLE OF CYCLIC NUCLEOTIDES |
|---|
Cyclic AMP and cyclic GMP are second messengers that mediate the physiologic responses to a host of hormones, neurotransmitters, autacoids, and drugs. The basic principles by
which these second messengers act are illustrated in Figure 2,
and they are identical regardless of the biologic response (for
review see reference 69). Briefly, cyclic AMP and cyclic GMP
are formed from ATP and GTP by the catalytic action of adenylyl cyclase and guanylyl cyclase, respectively. These enzymes are activated either directly (e.g., forskolin for adenylyl
cyclase and nitric oxide for guanylyl cyclase) or indirectly
through cell-surface receptors, e.g., 
-adrenoceptor agonists
and prostaglandin E2 (PGE2) for adenylyl cyclase and atriopeptins for guanylyl cyclase. As the intracellular concentrations of the cyclic nucleotides rise they bind to and activate
their target enzymes, PKA and protein kinase G (PKG).
These protein kinases phosphorylate substrates (e.g., ion
channels, contractile proteins, transcription factors) that regulate key cellular functions. Phosphorylation alters the activity
of these substrates, and thus, changes cellular activity. As cyclic nucleotides are inactivated by PDEs there is a corresponding decrease in protein kinase activity. Phosphoprotein phosphatases dephosphorylate substrates, and cellular activity
returns to normal. Obviously, altering the rate of cyclic nucleotide formation or degradation will change the activation state
of these pathways.
Immune and Inflammatory Cells
Cyclic nucleotides, particularly cyclic AMP, have important regulatory roles in virtually all cell types involved in the pathophysiology of asthma (Figure 4). Of paramount importance is the observation that, in general, cyclic AMP broadly suppresses the activity of immune and inflammatory cells (16, 41, 70). In contrast to the role of cyclic AMP, little evidence supports a substantive role for cyclic GMP in inflammatory cells.
|
Paradoxically, there are cell types in which cyclic AMP appears to exert both inhibitory and stimulatory influences. For
example, whereas cyclic AMP inhibits IFN
production from
mitogen-stimulated peripheral blood lymphocytes or antigen-stimulated T cell lines (74), it acts synergistically with Ca2+ to
increase IFN production in cytolytic T cells (75). Moreover, cyclic AMP inhibits B cell proliferation induced by IL-2, but it
stimulates proliferation induced by IL-4 (76). Conflicting evidence also exists regarding the effect of cyclic AMP on immunoglobulin synthesis by B lymphocytes (77, 78). On a molecular level such seemingly dichotomous effects can be explained
in part by the pleiotropic actions of a family of PKA-activated
transcription regulators called cyclic AMP response element-binding (CREB) proteins (79). Once activated, these proteins
bind to cyclic AMP response elements of various genes and
act either as positive regulators of transcription or, less frequently, as dominant-negative inhibitors. Because the complement of CREB proteins differs among cell types and because
their effects are dependent upon the stimulus, CREB activation can produce a number of apparently conflicting effects.
Airway Smooth Muscle
Overwhelming evidence indicates that both cyclic AMP and
cyclic GMP mediate relaxation of airway smooth muscle via
"classic" protein kinase/protein phosphorylation cascade mechanisms (80) (Figure 2). Of importance to asthma therapeutics
is the belief that -adrenoceptor agonists produce bronchodilation by elevating cyclic AMP content and activating PKA. However, recent discoveries have added new twists to the dogma.
First, the prevailing view that cyclic AMP relaxes airway smooth
muscle exclusively by activating PKA is challenged by evidence
that cyclic AMP may act via PKG as well (80). Even more heretical is the possibility that -adrenoceptor agonists produce
bronchodilation at least partly via a cyclic AMP-independent
mechanism, i.e., through direct regulation of Ca2+-dependent
K+ channels (80, 85, 86). As intriguing as these proposals are
they do not overturn the fundamental precept that cyclic AMP is a second messenger that mediates airway smooth muscle relaxation. They simply suggest that PKA may not be the sole
molecular target for cyclic AMP and that -adrenoceptor agonists may act by both cyclic AMP-dependent and cyclic-AMP-independent mechanisms.
Cyclic AMP may have an additional role in modulating airway smooth muscle hypertrophy and hyperplasia, which are common morphologic features of chronic asthma (87, 88). The role of cyclic AMP in regulating cellular proliferation is extremely complex in that it produces a cytostatic effect in early G0 to G1 transition and mid-G1 phase in many cell types (89, 90), but promotes the transition from G1 to S phase in others (91). Regardless, a number of studies support the proposal that cyclic AMP exerts an overall inhibitory influence on airway smooth muscle proliferation (92).
Epithelial Cells, Endothelial Cells, and Nerves
Epithelial cells protect the lung from the external environment, transport electrolytes to maintain ionic homeostasis of the airway surface, and provide a mucociliary transport system to rid the airway of cell debris and particulate matter. The airway epithelium is also an important source of bronchodilatory autacoids (96) as well as inflammatory cytokines and chemokines (97). Little is known about the role of cyclic AMP beyond its well-defined effect on ion transport (98, 99) and its ability to increase and synchronize ciliary beat frequency (100). Cyclic GMP has the opposite effect on the function of cilia (101).
Vascular endothelial cells act as gatekeepers to prevent or, upon receiving the appropriate stimulus, promote the infiltration of inflammatory cells and plasma proteins into the airway. Increases in either cyclic AMP or cyclic GMP content in endothelial cells are linked to a reversal of microvascular hyperpermeability induced by a number of insults (102). Moreover, cyclic AMP inhibits the expression of ELAM-1 and VCAM, adhesion molecules that play a central role in the recruitment of inflammatory cells into the airway (105).
Limited information is available concerning the role of cyclic nucleotides in airway neurotransmission. Recent studies
employing direct measurement of neurotransmitter release
from isolated tracheal preparations indicate that -adrenoceptor agonists augment acetylcholine release (106, 107). This effect is mimicked by forskolin and 8-bromo-cyclic AMP (106),
suggesting that the classic cyclic AMP second messenger cascade is involved. The role of cyclic nucleotides in regulating
nonadrenergic noncholinergic (NANC) neuronal activity in
the airway has not been studied directly, although, as discussed later, PDE inhibitors indeed modulate the activity of
these nerves.
| |
EFFECTS OF ISOZYME-SELECTIVE PHOSPHODIESTERASE INHIBITORS IN VITRO |
|---|
Immune and Inflammatory Cells
An appreciation of the functional role of individual PDE isozymes in various cells and tissues requires critical analysis of results from two distinct experimental approaches. The first of these approaches involves using classic biochemical and pharmacologic techniques to define the PDE isozyme profile of cell types of interest. As summarized in Table 4, information is available on the PDE isozyme profile of most immune and inflammatory cells that are implicated as central players in the pathophysiology of asthma. Of course, the mere presence of a particular isozyme in a cell extract does not provide convincing evidence that the isozyme is important in regulating cyclic nucleotide content and, in turn, cell function. To address this issue a second experimental approach is employed. This involves evaluating the biochemical and functional effects of isozyme-selective PDE inhibitors in intact tissues. The results of these studies are summarized below and in Table 5. The overriding conclusion is that PDE4 and, to a lesser extent, PDE3 represent the major cyclic AMP metabolizing enzyme families in all immunocompetent cells. Moreover, the functional data suggest that PDE4 is a primary molecular target for novel antiasthmatic drugs.
|
|
Basophils and mast cells. Human basophils contain PDE3, PDE4, and PDE5, with the first two isozymes accounting for virtually all cellular cyclic AMP hydrolytic activity (108). Phosphodiesterase 4 inhibitors suppress antigen-induced release of histamine (108) and the cysteinyl leukotrienes (LT) (108, 111). Despite the presence of PDE3 and PDE5 in basophils, inhibition of these enzymes fails to decrease mediator release (108, 109). PDE3 inhibitors do, however, potentiate the effects of PDE4 inhibitors (108).
Perhaps because isolation of large quantities of human mast cells is technically challenging, the PDE isozyme profile in these cells awaits detailed evaluation. However, preliminary experiments on enriched human lung mast cells suggest that these cells contain both PDE3 and PDE4 (111, 112). Consistent with this finding are reports that both PDE3 and PDE4 inhibitors, but not a PDE5 inhibitor, reduce antigen-driven histamine release from human lung (113, 114) and skin (114) mast cells. These results contrast with those of another study in which neither PDE3 nor PDE4 inhibitors suppressed mediator release from human lung mast cells (111).
B Lymphocytes. Surprisingly little is known about the PDE profile of B cells or the action of PDE inhibitors on IgE secretion. The limited work conducted thus far indicates that PDE4 accounts for nearly all of the cyclic AMP hydrolyzing activity of Jijoye cells, a human B cell line (115). The effects of PDE inhibitors on B lymphocyte function have not been analyzed in detail, although the nonselective inhibitor IBMX profoundly inhibits spontaneous IgE secretion from LA350 cells, a transformed human B cell line (116). The effect of IBMX on mixed blood mononuclear cells from atopic subjects is more complex, with low concentrations of the PDE inhibitor enhancing IgE release and greater concentrations inhibiting release (117). Ro 20-1724, a PDE4 inhibitor, reduces spontaneous release of IgE from mononuclear cells isolated from atopic subjects (118). Because this inhibitory effect is not observed with purified B cells it is likely to be an indirect effect, perhaps through inhibition of IL-4 secretion from T cells (119).
Eosinophils. Both human (120, 121) and guinea-pig (122,
123) eosinophils contain one or more membrane-bound and
cytosolic forms of PDE4 that account for virtually all of the
cyclic AMP metabolizing capacity of the cell. Consistent with
this isozyme profile, PDE4 inhibitors suppress a broad spectrum of eosinophil functions. In guinea-pig eosinophils these
compounds inhibit basal superoxide anion production (122)
as well as superoxide production stimulated by N-formyl-
methionyl-leucyl-phenylalanine (fMLP) (124, 125) and LTB4
(126). Other functions of guinea-pig eosinophils inhibited by
this compound class include: (1) major basic protein and eosinophil cationic protein release (126), (2) cytokine-driven adhesion to endothelial cells (125), (3) platelet-activating factor
(PAF)-induced and C5a-induced aggregation (127), (4) leukotriene (LT) B4 generation (128), (5) opsonized zymosan-
induced H2O2 production (123), (6) phorbol-ester- and TNF
-induced adhesion to endothelial cells (125), and (7) C5a- and
LTB4-induced chemotaxis (129). Studies with human eosinophils are less plentiful, although the available data suggest that
the activity of PDE4 inhibitors in these cells is broadly similar
to that observed with guinea-pig eosinophils. Thus, rolipram
and CP-80633, a PDE4 inhibitor, block opsonized zymosan-stimulated superoxide generation (121, 129). Rolipram also
inhibits PAF- and C5a-induced LTC4 production and chemotaxis (130, 131), as well as PAF-induced expression of CD11b
(132). In the presence, but not in the absence, of a -adrenoceptor agonist both rolipram and RP 73401, a newer PDE4 inhibitor, reduce the release of eosinophil cationic protein and
eosinophil-derived neurotoxin in response to C5a (120).
In contrast to the consistent suppressant effect of PDE4 inhibitors on eosinophil function, inhibitors of PDE3 or PDE1 have no effect (121, 125, 128).
Hypothetically, PDE inhibitors and activators of adenylyl
cyclase should act synergistically if cyclic AMP is modulating
the biologic response in question. The data supporting this tenet are conflicting with regard to eosinophil function. On one
hand, PDE4 inhibitors act synergistically with adenylyl cyclase
activators (i.e., salbutamol, PGE1) to inhibit aggregation of
guinea-pig eosinophils (127) and the secretion of granule constituents from human eosinophils (120). On the other hand,
salbutamol and rolipram do not act synergistically to inhibit
LTC4 synthesis (131) or superoxide generation (121) in human
eosinophils. As with airway smooth muscle (80, 85), these results might indicate that cyclic-AMP-independent mechanisms contribute to the ability of -adrenoceptor agonists to
inhibit certain eosinophil functions (121).
Macrophages and monocytes. Human monocytes contain
predominantly PDE4 (68, 133, 134) along with a small amount
of PDE3 (135). Maturation of monocytes into macrophages
markedly alters the PDE isozyme profile. In human alveolar
macrophages the activity of PDE3 increases nearly to the level
of PDE4 (133). PDE1 and, to a lesser extent, PDE5 are also
present in macrophages (133), thus giving these cells the capacity to hydrolyze both cyclic AMP and cyclic GMP. The most
striking action of PDE inhibitors on monocyte function is their
ability to inhibit lipopolysaccharide-induced TNF production
(128, 129, 135). In contrast, inhibitors of PDE1, PDE3, and
PDE5 are either considerably less active or devoid of activity
(135, 138, 139, 143). The effects of PDE4 inhibitors are not
limited to regulation of TNF generation, as they also suppress
fMLP-induced arachidonic acid release (144) as well as calcium-ionophore-induced LTB4 and LTC4 generation (145). Interestingly, despite the inhibitory effect of these compounds on
arachidonic acid release and LT production, they have no effect on the generation of prostanoids (145). Indeed, several
other monocyte functions are not regulated by PDE4 inhibitors. In particular, PDE4 inhibitors have little or no effect on
the production of IL-1, IL-6, superoxide anion or nitric oxide
(138, 139, 143, 146, 147).
Studies to evaluate interactions between PDE4 inhibitors
and adenylyl cyclase inhibitors have also been conducted. With
regard to inhibition of TNF elaboration, PDE4 inhibitors act
in an additive rather than a synergistic manner with -adrenoceptor agonists (135) or the prostacyclin analog, cicaprost
(134, 140). Interestingly, however, PDE4 inhibitors act synergistically with PGE2 to reduce TNF secretion under the same
conditions that they fail to synergize with -adrenoceptor agonists (135, 140). Although puzzling at first glance, these data
might indicate that -adrenoceptor and prostacyclin receptor
agonists suppress TNF formation via a mechanism independent of cyclic AMP (135). The actions of PGE2, in contrast, may be mediated solely by standard cyclic AMP-dependent
mechanisms.
Because of the technical difficulties associated with obtaining human alveolar macrophages, information on the actions
of PDE inhibitors in these cells is less plentiful than is information on peripheral blood monocytes. Nonetheless, consistent with the PDE isozyme profile of alveolar macrophages,
either rolipram or motapizone, a PDE3 inhibitor, reduce but
do not abolish LPS-induced TNF production (148). The combinaton of both inhibitors is more effective than the individual
agents (148), supporting the concept that both PDE3 and
PDE4 are functionally important in the macrophage. In contrast, Ro 20-1724, the first PDE4 inhibitor described, does not
inhibit the production of thromboxane B2 or superoxide anion in response to opsonized zymosan (149). Thus, as with monocytes, the actions of PDE inhibitors are not universal but instead depend upon the function being evaluated. A study with
murine macrophages indicates that rolipram increases IL-10
synthesis (150). This prompts speculation that a portion of rolipram's inhibitory effect on murine TNF synthesis is mediated indirectly through IL-10, which itself suppresses TNF
production (150).
Neutrophils. PDE4 is the overwhelmingly predominant cyclic-AMP-metabolizing enzyme in human neutrophils (151-
153). Also present in these cells is a small amount of PDE5
(151, 153). In peripheral blood neutrophils, PDE4 inhibitors
reduce superoxide anion production in response to a number
of stimuli, including fMLP (151), C5a (152), granulocyte-macrophage colony-stimulating factor (155) and TNF (156).
In contrast, phagocytosis-induced respiratory burst is not reduced by PDE4 inhibitors (152). In addition, suppression of
fMLP-induced superoxide formation in tissue neutrophils is somewhat less susceptible to PDE4 inhibitors than blood neutrophils (154). Neutrophil functions other than superoxide formation are also regulated by PDE4 inhibitors, as compounds
of this class reduce fMLP-induced degranulation (141, 157,
158), LT biosynthesis (153) and adhesion to endothelial cells
(159). The latter effect is probably due to the ability of PDE4
inhibitors to reduce the expression of neutrophil CD11b/
CD18 (132, 159). Adenylyl cyclase activators and PDE4 inhibitors act synergistically to suppress most, but not all (132),
neutrophil functions (151, 155, 157, 159).
T Lymphocytes. Initial studies with mixed T cells indicated the presence of roughly equal amounts of cytosolic PDE4 and membrane bound PDE3 (160, 161). Recently, studies with T lymphocyte cell lines (40) as well as with isolated CD4+ and CD8+ lymphocytes (162, 163) suggest the presence of a third cyclic-AMP-metabolizing enzyme, PDE7. In addition, both CD4+ and CD8+ cells contain small amounts of PDE1, PDE2, and PDE5 (162).
The first comprehensive study of the effects of isozyme- selective PDE inhibitors on T cell function was conducted by Robicsek and colleagues (160). Used alone, both Ro 20-1724 and CI-930, a PDE3 inhibitor, partially inhibit phytohemagglutinin (PHA)-driven T cell proliferation. Combining the two agents results in an inhibitory effect that is at least additive. The functional results are thus consistent with the presence of both PDE3 and PDE4 in these cells. Interestingly, however, the magnitude of the inhibitory effect of a combination of Ro 20-1724 and CI-930 is less than that produced by papaverine (160), a nonselective PDE inhibitor. Perhaps the additional efficacy of nonselective PDE inhibitors signals a functional role of PDE7 as well as PDE3 and PDE4. Subsequent studies largely confirmed the initial results in that PDE4 inhibitors reduce proliferation of peripheral blood mononuclear cells induced by PHA (163), specific antigen (164, 166), or a combination of phorbol ester and an anti-CD3 antibody (163, 167). In contrast to the earlier study, however, results from the latter studies suggest that PDE4 is considerably more important than PDE3 as an antiproliferative target (163). Despite the fact that PDE3 inhibitors have little effect on their own, they accentuate the actions of PDE4 inhibitors under certain conditions (163). Studies with purified monocytes and T cells suggest that the ability of PDE4 inhibitors to down-regulate proliferation in mixed cell populations is predominantly due to a direct effect on lymphocytes (168).
Interestingly, proliferation of peripheral blood mononuclear cells from atopic subjects is more sensitive to PDE4 inhibitors than is proliferation of cells from nonatopic subjects (165, 169). This phenomenon may stem in part from the fact that PDE4 isolated from atopic leukocytes is inherently more sensitive to inhibition by a variety of PDE4 inhibitors (170).
The mechanism whereby cyclic AMP and PDE inhibitors suppress T cell proliferation is likely to be complex. One potential target, however, is regulation of IL-2 synthesis. Indeed, agents that increase cyclic AMP content, including PDE4 inhibitors, reduce the expression of this cytokine (163, 171). However, the suppressant effect of rolipram on T cell proliferation is not overcome by replacement of IL-2 in culture medium (164). This suggests that at least a portion of rolipram's inhibitory effect on lymphocyte proliferation occurs at sites distal to IL-2 generation. This proposal is consistent with the view that cyclic AMP can inhibit IL-2-stimulated T cell proliferation by suppressing G1 progression (174, 175).
The production of several cytokines in addition to IL-2 is
modulated by PDE inhibitors. In peripheral blood mononuclear cells, PDE4 inhibitors suppress the production of IFN,
granulocyte-macrophage colony-stimulating factor, and IL-5
in response to antigen (158, 166) or mitogen (173). Similar to
the results on T cell proliferation, PDE3 inhibitors have no effect on cytokine generation on their own, but they enhance
the effect of PDE4 inhibitors (166). Phosphodiesterase 4 inhibitors also reduce the production of the above-mentioned cytokines (167) as well as IL-13 (176, 177) from antigen-stimulated T cell clones derived from atopic donors. The effects of
PDE4 inhibitors on IL-4 production are somewhat variable.
These agents suppress anti-CD3-stimulated IL-4 production
in peripheral mononuclear cells (119), but they do not inhibit
antigen-driven IL-4 production in these cells (166). No such
ambiguity exists in purified T cells challenged with antigen
(168) or Th2 cell clones stimulated with antigen (176, 177),
anti-CD3, or mitogen (167). In these settings PDE4 inhibitors
simultaneously reduce the generation of IL-4, IL-5, and IL-13.
In contrast to observations made using mixed cell preparations, PDE3 inhibitors have no effect on cytokine production
from T cell clones (176, 177). Overall, human Th1 and Th2
cells appear to be equally responsive to the antiproliferative and cytokine-suppressing effects of PDE4 inhibitors (166,
176, 177).
Airway Smooth Muscle
Tone. The PDE isozyme profile of human airway smooth muscle is complex. Human trachealis (178) and peripheral airways (179) contain PDEs 1, 2, 3, 4, and 5. Consistent with the presence of large amounts of PDE3 and PDE4, selective inhibitors of either isozyme partially reverse spontaneous tone of human isolated bronchi (180, 181), with PDE3 inhibitors being somewhat more effective (180). Studies on tissues precontracted with exogenously added spasmogens confirm and extend these conclusions (178, 179). Zaprinast, a PDE5 inhibitor, has a modest relaxant effect against spontaneous tone (180) but virtually no effect against agonist-induced tone (178).
Functional analyses of PDE isozymes suggest that PDE3 and PDE4 coregulate cyclic AMP content in human airway smooth muscle. This appears to be the case as either a combination of PDE3 and PDE4 inhibitors or dual PDE3/4 inhibitors produce a much greater bronchorelaxant effect than individual isozyme-selective agents alone (178, 179).
The activation state of adenylyl and guanylyl cyclase influences the bronchorelaxant efficacy of PDE inhibitors. Thus,
-adrenoceptor agonists potentiate the actions of PDE3 and
PDE4 inhibitors in canine trachealis (182, 183), and sodium
nitroprusside, an activator of guanylyl cyclase, potentiates the
effects of PDE5 inhibitors (184). The scenario is less clear in
human airway, where the synergistic interaction between PDE
inhibitors and exogenously added cyclase activators is small
and variable (178). This equivocal effect could be related to
cyclic-AMP-independent effects of -adrenoceptor agonists in
human airway smooth muscle (80, 178). Interestingly, however, the ability of PDE3 and PDE4 inhibitors to relax the human isolated airway is highly dependent upon the presence of endogenous prostanoids (178). Thus, an increase in basal cyclic AMP synthesis might be required for the full expression of
the bronchorelaxant activity of PDE3 and PDE4 inhibitors.
Mitogenesis. The actions of isozyme-selective PDE inhibitors on airway smooth muscle proliferation have yet to be reported. However, Souness and coworkers (185) carried out an in-depth analysis of the role of PDE isozymes in modulating mitogen-stimulated proliferation and [3H]thymidine incorporation, an index of mitogenesis, in porcine aortic smooth muscle cells. Mitogenesis induced by fetal bovine serum is reduced only modestly by either PDE3 or PDE4 inhibitors. In contrast, combining a PDE3 inhibitor with a PDE4 inhibitor has a profound inhibitory effect on proliferation, consistent with the proposal that these PDEs coregulate cyclic AMP content in vascular smooth muscle. Zaprinast has no effect on mitogenesis in pig aortic smooth muscle cells, either alone or in combination with a guanylyl cyclase activator, suggesting no role for PDE5 or cyclic GMP in modulating proliferation. Clearly, there is a need to conduct similar studies on airway smooth muscle.
Epithelial Cells, Endothelial Cells, and Nerves
A preliminary report indicates that several PDE isozymes exist in primary cultures of human airway epithelial cells (186). Both PDE3 and PDE4 are present, although PDE4 is more
prominent. Caution should be used in interpreting these results as this preparation is likely to contain multiple cell types.
Despite the apparent predominance of PDE4 in primary airway epithelial cell cultures, inhibitors of PDE3, but not of
PDE4, regulate Cl
conductance in a human airway epithelial
cell line (Calu-3) (187). The nonselective inhibitor 3-isobutyl-1-methylxanthine inhibits bradykinin-stimulated production of
PGE2 in human primary epithelial cells (186). Effects of PDE
inhibitors on chemokine generation from airway epithelial
cells have not been reported.
Human umbilical vein endothelial cells contain large amounts of PDEs 2, 3, and 4 (188). Increased permeability of human endothelial cell monolayers induced by thrombin or Escherichia coli hemolysin is blocked by inhibitors of PDE3 or PDE4, as well as by dual PDE3/4 inhibitors (188). Consistent with the inhibitory action of cyclic GMP on endothelial cell permeability and the role of PDE2 as the major cyclic GMP-metabolizing enzyme in these cells, H2O2-induced hyperpermeability of porcine pulmonary artery endothelial cells is reversed by EHNA, a selective PDE2 inhibitor (104).
The primary relaxant innervation to the airway is provided by inhibitory NANC (iNANC) nerves (189). The transmitter for this relaxant response is NO (190), which relaxes airway smooth muscle by stimulating guanylyl cyclase and activating the protein kinase G cascade (193). PDE4 inhibitors, but not PDE3 inhibitors, potentiate iNANC-induced relaxation of human bronchial smooth muscle (31) and guinea-pig trachea (196). Because PDE4 does not metabolize cyclic GMP, the most obvious explanation for these results is that PDE4 inhibitors potentiate NO generation presynaptically. However, this may not be the case as PDE4 inhibitors also potentiate the relaxant responses to nitrovasodilators (31), agents that relax airway smooth muscle directly via a cyclic-GMP-mediated mechanism. One hypothetical explanation for this is that by virtue of their ability to elevate cyclic GMP content, nitrovasodilators indirectly inhibit PDE3 (cyclic GMP is a competitive inhibitor of PDE3-mediated cyclic AMP hydrolysis). Thus, in the presence of a nitrovasodilator and a PDE4 inhibitor, both cyclic-AMP-metabolizing enzymes (i.e., PDE3 and PDE4) are inhibited, resulting in an increase in cyclic AMP content and relaxation (31). Interestingly, zaprinast, a PDE5 inhibitor, has no effect on iNANC relaxations of human bronchi (31), even though it potentiates relaxation induced by NO-donors (31, 178).
Excitatory NANC (eNANC) contractions of guinea-pig isolated bronchi are reduced by PDE4 inhibitors and, to a lesser extent, by PDE3 inhibitors (197). Inhibition of PDE5 has no effect on eNANC contractions (197). That the inhibitory actions result from reduced transmitter (i.e., tachykinins) release rather than functional antagonism at the level of the smooth muscle is supported by the fact that neither PDE3 nor PDE4 inhibitors abrogate capcaicin-induced or neurokinin-A-induced contractions (197). In contrast to the marked effects of PDE4 inhibitors on NANC activity in the airway, these compounds do not influence cholinergic neurotransmission in this tissue (31, 197). Interestingly, however, vinpocetine, a PDE1 inhibitor, reduces electrical-field-stimulation-induced acetylcholine release from guinea-pig trachea (198). This result hints that PDE1 regulates cyclic AMP content, cyclic GMP content, or both in cholinergic neurons of the airway, and that one or both of these cyclic nucleotides inhibits acetylcholine release.
| |
ACTIONS OF ISOZYME-SELECTIVE PHOSPHODIESTERASE INHIBITORS IN VIVO |
|---|
Anti-inflammatory Activity
The effects of isozyme-selective PDE inhibitors in models of pulmonary inflammation have been evaluated in a large number of studies (Table 6). The overriding theme that emerges from these studies is that PDE4 inhibitors are active in a wide spectrum of pulmonary inflammation models.
|
As might be predicted from the actions of PDE4 inhibitors on mast cell degranulation, pretreatment with these agents inhibits antigen-induced bronchoconstriction in guinea pigs (199), rabbits (205), and cynomolgus monkeys (206). That this effect is primarily due to an inhibition of mast cell degranulation rather than to a direct bronchodilator effect is suggested by the fact that doses of PDE4 inhibitors that virtually abolish the response to antigen do not prevent bronchoconstriction induced by histamine or LTD4 (199, 200, 203). Although most studies indicate that PDE4 inhibitors reduce antigen-induced bronchoconstriction, this effect was not observed in at least two studies in which different immunization procedures and/or dosing regimens were used (207, 208). In contrast to the actions of PDE4 inhibitors, PDE3 and PDE5 inhibitors have little or no effect on the early response to antigen (199, 203).
The most impressive property of PDE4 inhibitors in models of pulmonary inflammation is their ability to abolish antigen-driven eosinophil infiltration. This effect occurs in several species, including guinea pig (199, 208, 209), rat (210, 211), rabbit (205, 207), and monkey (206). The most straightforward explanation for this action is that PDE4 inhibitors ablate mast cell degranulation and, consequently, reduce the generation of chemotactic mediators that are responsible for eosinophil recruitment. Two lines of evidence argue against this as being the sole mechanism. First, PDE4 inhibitors reduce eosinophil infiltration even when they are administered after the antigen challenge (212, 213), i.e., after mast cell degranulation has already taken place. Second, PDE4 inhibitors suppress eosinophil influx induced by a number of agents that produce chemotaxis directly in a mast-cell-independent manner (204, 213). These data suggest that PDE4 inhibitors reduce eosinophil influx by at least two general mechanisms, one that involves an inhibition of mediator release and a second that involves a more generalized inhibitory effect on inflammatory cell trafficking. In addition to their ability to suppress eosinophil infiltration, PDE4 inhibitors also reduce the activation of these cells in vivo as assessed by decreased eosinophil peroxidase release into bronchoalveolar (214, 216) or pleural (204) lavage fluids.
Similar to their effects on eosinophil influx, PDE4 inhibitors markedly reduce antigen-stimulated neutrophil infiltration in the guinea pig (208, 217), rat (210, 211), and monkey (206). On the other hand, these compounds do not inhibit antigen-induced pulmonary neutrophilia in the rabbit (207). PDE4 inhibitors also fail to reduce neutrophil influx into the rat airway in response to lipopolysaccharide (218, 219).
In contrast to the observation that PDE4 inhibitors generally suppress inflammatory cell infiltration into the airway, inhibitors of PDE1 (209), PDE3 (208, 209, 214), or PDE5 (214, 215) are, with little exception (211), without effect.
Additional beneficial effects of PDE4 inhibitors in vivo include their abilities to reduce airway hyperreactivity (199, 202, 206, 217, 220) and pulmonary microvascular leakage (202, 218, 221, 222) induced by a number of challenges and in a number of species. In contrast to the activity of PDE3 inhibitors on isolated endothelial cells, these compounds do not inhibit microvascular leakage induced by histamine in guinea pigs (221). On the other hand, they reduce pulmonary microvascular leakage in response to lipopolysaccharide challenge in the same species (219) and in rats (218). Because lipopolysaccharide is likely to produce its effects indirectly by stimulating the release of mediators from alveolar macrophages, it is possible that PDE3 inhibitors suppress the action of these cells rather than having a direct effect on the pulmonary endothelium. Although PDE3 and PDE4 inhibitors act in an additive fashion in selected immunocompetent cells (see above), such an interaction has yet to be demonstrated convincingly in vivo (201, 203, 210, 216).
Factors Influencing the Anti-inflammatory Activity of PDE Inhibitors
Two critical factors influence the anti-inflammatory efficacy
of PDE inhibitors in vivo. The first relates to the potential synergistic interaction between the anti-inflammatory activities of
PDE inhibitors and endogenous activators of adenylyl cyclase (e.g., catecholamines, PGE2) (16, 70). As discussed earlier many, albeit not all, actions of PDE inhibitors on a number of inflammatory cell functions in vitro are potentiated by adenylyl cyclase activators. Several recent studies in in vivo models
mirror the results obtained from in vitro experiments. For example, PDE4 inhibitors act synergistically with circulating epinephrine to suppress antigen-induced mast cell degranulation
in anesthetized guinea pigs (223). Under identical conditions,
however, circulating catecholamines do not influence the ability of PDE4 inhibitors to abrogate antigen-induced pulmonary
eosinophilia (223) or IL-5-induced pleural eosinophilia (204).
In the mouse, endogenous catecholamines act in conjunction
with PDE4 inhibitors to elevate plasma cyclic AMP levels
(224) and greatly augment the ability of PDE4 inhibitors to reduce arachidonic-acid-induced cutaneous inflammation (145). Conversely, catecholamines are not required for the PDE4 inhibitors to suppress lipopolysaccharide-induced TNF production in the mouse (224). Although synergistic anti-inflammatory interactions between PDE inhibitors and PGE2 or
prostacyclin occur in vitro (135, 140), a role for endogenous
prostanoids in potentiating the effects of PDE inhibitors in
vivo has not been demonstrated (145, 223). Regardless, these
data suggest that studies conducted using isolated cells underestimate the activity of PDE inhibitors in in vivo settings
where endogenous activators of adenylyl cyclase are present.
A second factor to consider when attempting to predict the therapeutic activity of PDE inhibitors is the likelihood that these compounds impact inflammatory processes at multiple points. The role of PDE4 inhibitors in regulating eosinophil function is a case in point (Figure 5). As described earlier, these compounds have a multidimensional inhibitory effect on eosinophil function. Thus, PDE4 inhibitors can modify the role of the eosinophil in pathologic processes by: (1) suppressing the release of chemotactic mediators, (2) inhibiting chemotaxis directly, (3) reducing endothelial cell adhesion, (4) inhibiting IL-5 secretion, and (5) inhibiting the generation and release of eosinophil-derived mediators and cytotoxic substances. It is important to recognize that the pleiotropic action of PDE4 inhibitors is not limited to effects on eosinophils. Instead, it is likely that scenarios similar to the one depicted in Figure 5 are applicable to virtually all cells involved in airway inflammation.
|
Bronchodilatory Activity
The bronchodilator activity of isozyme-selective PDE inhibitors has been assessed in a number of studies. PDE3 inhibitors reverse 5-hydroxytryptamine-induced bronchoconstriction in dogs (225) and improve baseline pulmonary function in the same species (226). Phosphodiesterase 3 inhibitors also produce bronchodilation in anesthetized guinea pigs (199, 203, 227, 228). Predictably, substantial changes in cardiovascular function accompany the bronchodilation induced by PDE3 inhibitors. Phosphodiesterase 4 inhibitors reverse 5-hydroxytryptamine-induced bronchoconstriction in the dog (225, 229) and histamine-induced bronchoconstriction in the guinea pig (228). In contrast, PDE4 inhibitors are virtually inactive as bronchodilators when administered prophylactically (200, 203, 220), but act synergistically with PDE3 inhibitors to prevent LTD4 or histamine-induced bronchospasm (203, 230). This observation is consistent with the proposal that PDE3 and PDE4 coregulate cyclic AMP content in airway smooth muscle. Zardaverine, a dual PDE3/4 inhibitor, produces a bronchodilatory effect similar to that produced by the coadministration of PDE3 and PDE4 inhibitors (203). Regrettably, any advantage offered by the increased bronchodilatory activity of dual PDE3/4 inhibitors is offset by the profound cardiovascular actions of these compounds (203).
SK&F 96231, a PDE5 inhibitor, reverses bronchoconstriction in the guinea pig induced by infusion of U46619 (226) and prevents bronchoconstriction induced by histamine (230).
On balance, the data suggest that both PDE3 and PDE4 inhibitors possess moderate bronchodilator activity in animals, with the former class of compound being somewhat more effective than the latter.
| |
REGULATION OF PDE ACTIVITY |
|---|
Mechanisms
Alterations in the amount or activity of PDEs within effector cells profoundly affect their responsiveness to endogenous and exogenous stimuli. Two general types of PDE regulation occur (for reviews see references 10, 11, and 17). The first type, designated "short-term regulation," involves activation of second messenger pathways and the consequent allosteric modulation of enzyme activity. The second general mechanism, called "long-term regulation," occurs through increased enzyme synthesis. Several PDE families are subject to one or both forms of regulation. This section, however, focuses on the long-term regulation of PDE4, a process that is proposed to have important implications in the pathogenesis and treatment of asthma (54, 231).
Observations made as long ago as the early 1970s indicate that treatment of various cells and tissues with agents that increase cyclic AMP content stimulate an increase in PDE activity (for discussion see reference 232). This process appears to be a universal homeostatic mechanism whereby target cells regulate their responsiveness to hormones and autacoids that activate adenylyl cyclase by generating a reciprocal increase in PDE activity. Unlike short-term regulation of PDE activity, which provides ephemeral, moment-to-moment modulation, long-term regulation of PDE activity develops over a protracted time frame and is only slowly reversible upon removal of the adenylyl-cyclase-stimulating agent. The molecular mechanism by which long-term regulation occurs was elucidated by Conti and coworkers (66, 67, 233, 234) who used rat Sertoli cells as a model system. These studies revealed that treating cells with cyclic-AMP-elevating agents results in increased expression of selected PDE4 subtypes. The regulation is very specific in that even within a single PDE4 subtype, the expression of only certain splice variants is regulated. Within the PDE4D subfamily, for example, a prolonged elevation of cyclic AMP content in intact cells increases the expression of PDE4D1 and PDE4D2 but not the longer splice variants (235). This enhanced expression is due to activation of a specific intronic promoter within the PDE4 gene, perhaps through the protein kinase A/CREB pathway (235).
The work on Sertoli cells laid the foundation for subsequent investigations into the regulation of PDE4 expression in immunocompetent cells. Thus, stimulating cyclic AMP accumulation increases PDE4 expression in macrophages (236), monocytes (53, 55), U937 cells (50, 54, 232), a human premonocytic cell line, and Jurkat cells (237), a human T-cell line. In all cases the increase in PDE4 catalytic activity produced by cyclic-AMP-elevating agents is due to an increase in gene expression. However, the specific PDE4 subtypes that are up-regulated varies depending on the cell type being examined, suggesting a differential regulation of PDE4 synthesis on the basis of cell-specific use of promoters. For example, assessment of steady-state message and protein expression demonstrate an up-regulation of PDE4A and PDE4B in U937 cells (54). A similar pattern is observed in human monocytes treated with a variety of cyclic-AMP-elevating agents, cyclic AMP analogues or lipopolysaccharide (53, 55). On the other hand, stimulating Jurkat cells with forskolin up-regulates PDE4D1 and PDE4D2 as well as PDE3 (237). Interestingly, a reciprocal regulation of the expression of PDE4 subtypes is frequently observed. Thus, although adenylyl cyclase activators increase the expression of transcripts for PDE4A and PDE4B in U937 cells, they appear to decrease PDE4D3 mRNA expression (54). In contrast, forskolin increases the expression of PDE4D1 and PDE4D2 in Jurkat cells at the same time it decreases the apparent expression of a novel splice variant of PDE4A. Regardless of this reciprocal regulation, the net result is that total cellular cyclic AMP PDE activity is invariably elevated by exposing cells to cyclic AMP elevating agents.
The greatest degree of PDE4 induction is generally observed when cells are treated simultaneously with a -adrenoceptor agonist and a PDE inhibitor (232, 236). Although
-adrenoceptor agonists up-regulate PDE4 activity on their
own, the increase is often modest and transient. The reason
for this is the self-regulating nature of the system. That is, as
-adrenoceptor agonists activate the PKA cascade to up-regulate PDE4 expression, the newly synthesized PDE4 tends to
reduce intracellular cyclic AMP content. The moderation of
cyclic AMP concentrations leads to a partial reversal of the induction process.
The question of how regulation of PDE4 activity impacts cell function has been addressed in studies using U937 cells as a model system. In these cells up-regulation of PDE4 induced by pretreatment with salbutamol causes a heterologous desensitization to adenylyl cyclase activators (54). For example, up-regulating PDE4 activity results in a substantial reduction in the ability of PGE2 to elevate cyclic AMP content and inhibit LTD4-induced Ca2+ mobilization. On the other hand, the sensitivity of these cells to PGE2 is normalized by the addition of rolipram, suggesting that the salbutamol-induced desensitization stems from an increase in PDE4 activity. The functional consequences of PDE4 up-regulation in monocytes (53) are similar to those observed in U937 cells (53).
Therapeutic Implications of PDE4 Regulation
Alterations in the expression of PDE4 activity could influence
both the treatment and the pathophysiology of asthma. With regard to the treatment of asthma, the use of -adrenoceptor
agonists to control symptoms could up-regulate PDE4 activity
in various cells in the airway (54, 238). If this were to occur it
could have implications with regard to the observation that
excessive use of -adrenoceptor agonists is linked to a deterioration of asthma control (239). A number of phenomena
have been proposed as contributors to this detrimental effect,
including: (1) masking a deterioration of disease status, (2) increasing the antigen load to the distal airways, (3) compromising the "protective" role of lung mast cells, and (4) down-regulating -adrenoceptors (242). An additional possibility
to consider is that excessive use of -adrenoceptor agonists
might up-regulate PDE4 activity (54, 238). As discussed earlier, up-regulating PDE4 activity in inflammatory cells and airway smooth muscle could reduce their responsiveness to endogenous activators of adenylyl cyclase. This, in turn, could
allow inflammatory processes and bronchoconstriction to proceed unchecked (54, 238, 245). A question that remains to be
answered is whether routine use of -adrenoceptor agonists can up-regulate PDE4 activity in a therapeutic setting. With
respect to this question, Cockroft and colleagues (246) reported that the chronic use of inhaled salbutamol increases
airway responsiveness to allergen and causes tolerance to the
protective effect of the -adrenoceptor agonist against allergen challenge, presumably by diminishing its ability to suppress mast cell mediator release. This tolerance to -adrenoceptor agonists could in part be mediated by an up-regulation
of PDE activity (54).
Regarding the pathophysiologic role of PDE4 up-regulation, Hanifin and colleagues (231, 247, 248) proposed that elevated PDE activity represents a central biochemical abnormality that accounts for the aberrant hormone sensitivity of
leukocytes in atopic diseases. Indeed, a number of immunocompetent cells isolated from patients with atopic disorders
have a diminished functional responsiveness to -adrenoceptor agonists (247, 249). Furthermore, this reduced hormone sensitivity is associated with an elevated activity of PDE4
(253). At least one study failed to detect an increase in
PDE4 activity in leukocytes from atopic subjects (133), perhaps because of differences in the methods used to isolate cells
and assess PDE activity. Regardless, an alteration in the expression of PDE4 could contribute to the pathophysiology of atopic disorders by altering the balance between endogenous
proinflammatory and anti-inflammatory pathways. Normally,
activators of adenylyl cyclase such as epinephrine, PGE2, and
prostacyclin act as endogenous anti-inflammatory agents (70,
256). Theoretically, an increase in PDE4 activity would lead to
a loss in sensitivity to these endogenous activators of adenylyl
cyclase, thus amplifying the inflammatory processes associated
with atopy.
| |
CLINICAL STUDIES |
|---|
Information from clinical trials with isozyme-selective PDE inhibitors is limited. Interestingly, the first compounds to be tested in humans were not PDE4 inhibitors, but rather the PDE5 inhibitor zaprinast and the PDE3 inhibitors enoximone and cilostazol. The clinical results with zaprinast were equivocal. Although oral administration of this compound reduced exercise-induced bronchoconstriction in adult asthmatics, it had no effect on histamine-induced bronchoconstriction in the same subjects (257) or on exercise-induced bronchoconstriction in children (258). Intravenous administration of enoximone improves pulmonary function in patients with chronic obstructive pulmonary disease (259), and oral administration of cilostazol produces both a bronchodilator and a bronchoprotective effect in normal subjects (260). Because of the major role of PDE3 in regulating cyclic AMP content in the cardiovascular system (261), however, the pulmonary effects of these compounds are accompanied by increases in heart rate and decreases in blood pressure (259) as well as headache (260). Tibenelast, a weak PDE4 inhibitor, slightly increased FEV1 in asthmatic subjects, although this effect was not statistically significant (262). Perhaps more interesting are the results with CDP840, a member of a new generation of highly potent and selective PDE4 inhibitors (263). In a double-blind, placebo-controlled trial, oral administration of CDP840 for 9.5 days significantly ablated the late asthmatic response to allergen, but it had no effect on the early asthmatic response or on histamine-induced bronchoconstriction (264). These results are tantalizing in that they suggest that CDP840 suppresses the late asthmatic response through an anti-inflammatory action rather than a direct bronchodilator effect. This is consistent with the proposed pharmacologic profile of PDE4 inhibitors.
The bronchodilatory activity of two dual PDE3/4 inhibitors, benafentrine and zardaverine, has also been evaluated in the clinic, with mixed results. Benafentrine, administered to normal volunteers by inhalation, produced bronchodilation, whereas it was inactive when administered intravenously or orally (265). In one study of asthmatic subjects inhalation of zardaverine produced modest bronchodilation (266), whereas it had no effect in another study of patients with chronic obstructive pulmonary disease (267).
The results of the limited clinical trials conducted to date have yet to prove the utility of isozyme-selective PDE inhibitors. Obvious factors that should be considered in weighing these results include the possibility that the compounds used had insufficient potency or poor pharmacokinetic profiles. Additional possibilities include the use of inappropriate patient populations, clinical end points, and treatment durations.
Notwithstanding the factors detailed above, dose-limiting side effects resulting in poor therapeutic indices represent the greatest barrier to the full exploitation of the potential therapeutic actions of the isozyme-selective PDE inhibitors evaluated thus far. These side effects are produced as extensions of the pharmacology of the various classes of inhibitors, i.e., the compounds inhibit one or more PDE isozymes in nontarget tissues. The implication of this statement is that although targeting inhibitors for specific isozymes has led to an improvement in functional specificity vis-a-vis the functional specificity of nonselective PDE inhibitors, it has yet to yield "magic bullets." Administering compounds by inhalation could reduce the frequency and severity of side effects, although this was not the case in the only clinical study conducted to address this issue (266). Moreover, the virtual absence of published data on the efficacy of inhaled isozyme-selective PDE inhibitors in animal models suggests that systemic exposure is required for full therapeutic activity.
Signature side-effect profiles are known for each class of isozyme-selective PDE inhibitor. PDE3 inhibitors produce a number of cardiovascular side effects (268, 269), including life-threatening arrhythmias in patients with compromised cardiac function (270). The cardiovascular activity of dual PDE3/4 inhibitors is likely to be even more marked since PDE4 inhibitors, although devoid of cardiovascular effects on their own, potentiate the cardiovascular actions of PDE3 inhibitors (271). Nausea and vomiting were the dose-limiting side effects of rolipram in clinical trials evaluating the antidepressant activity of the compound (272). Indeed nausea, vomiting, and gastric acid secretion are class effects of first generation PDE4 inhibitors (263, 273). It is likely that the emetic activity of PDE4 inhibitors is due primarily to an action in the CNS (277), although direct effects on the gastrointestinal tract (e.g., acid secretion) could contribute to this phenomenon (278).
Although information of dose-limiting side effects of several isozyme-selective PDE inhibitors is in hand, one could argue that the potential therapeutic activity of these compounds
at smaller doses has not been adequately evaluated. Nonetheless, the lack of strikingly positive data from clinical studies
with first-generation isozyme-selective PDE inhibitorsparticularly PDE4 inhibitors, the class perceived to have the greatest therapeutic potentialsuggest that improvements in their
side-effect profiles are needed before the full therapeutic utility of this compound class can be evaluated.
| |
SECOND-GENERATION PDE4 INHIBITORS |
|---|
The perception that side effects will ultimately limit the therapeutic utility of first-generation PDE4 inhibitors has prompted efforts to design second-generation compounds with improved therapeutic indices. These efforts are focused on one of two molecular approaches. One hinges on targeting one of two unique conformers of PDE4 (263, 275, 276) and the other focuses on selectively targeting PDE4 subtypes (46).
Regarding the former approac