512 Thorax 1991;46:512-523 New drugs review Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from inhibitors: new opportunities for the treatment of

Theodore J Torphy, Bradley J Undem

Theophylline has been a mainstay in the treat- osine 3',5'-monophosphate (cGMP) concen- ment of asthma for over 50 years. Its thera- trations in key cells concerned in the patho- peutic value stems from a combination of anti- physiology of asthma (figure). Firstly, both inflammatory and bronchodilator activities in cAMP and cGMP mediate airway smooth addition to its ability to increase diaphrag- muscle relaxation,9 10 so a rise in either of these matic contractility.14 Although the phar- second messengers in airway smooth muscle macological effects of have been should result in bronchodilatation. Secondly, studied extensively both in the laboratory and the ability of PDE inhibitors to increase the in the clinic, the molecular mechanisms res- cAMP content of inflammatory cells leads to ponsible for its activity in asthma remain ill an inhibition of cell activation.5 1112 In princi- defined. In fact, several cellular activities ple, then, PDE inhibitors should possess anti- probably contribute to its action, including inflammatory activity. cyclic nucleotide phosphodiesterase inhibi- Although it is useful in the treatment of tion,' 5 adenosine receptor antagonism,6 asthma, the value of theophylline is limited by stimulation of catecholamine release,7 and a a narrow and a wide range poorly understood ability to increase the num- of gastrointestinal, central nervous system, ber and activity of suppressor T lym- and cardiovascular side effects.`4 These side phocytes.8 effects are generally ascribed to its lack of Of theophylline's many activities, phos- selectivity. Of particular importance in this phodiesterase inhibition may be the most respect is theophylline's ability to antagonise

important. Hypothetically, at least two adenosine receptors and inhibit phosphodies- http://thorax.bmj.com/ therapeutically beneficial effects could result terase activity in inappropriate tissues (for from inhibition of phosphodiesterase activity example, gastrointestinal, central nervous sys- and the consequent rise in intracellular aden- tem, cardiovascular).2 13 The development of osine 3',5'-monophosphate (cAMP) or guan- drugs that couple the efficacy of theophylline

Cyclic AMP CyclicGMP| on September 30, 2021 by guest. Protected copyright.

Hormone Hormone

Receptor NVD or EDRF Receptor

Adenylate PDE PDE Guanylate Cyclase Cyclase

ATP o cAMP O 5'-AMP 5'-GMP cGMP GTP

Department of cA-PK f. cA-PK cG-PK | cG-PK Pharmacology, Smith (inactive) (active) (active) (inactive) Kline Beecham Pharmaceuticals, King of Prussia, Pennsylvania 19406 T J Torphy I I Division of Clinical Immunology, Johns INFLAMMATORY CELL ACTIVITY AIRWAY SMOOTH MUSCLE TONE Hopkins University School of Medicine, Baltimore, Maryland Role of cyclic nucleotides in thefunction of inflammatory cells and airway smooth muscle. See textfor details. PDE- 21224, USA phosphodiesterase; cA-PK-cAMP dependent protein kinase; cG-PK-cGMP dependent protein kinase; NVD- B J Undem nitrovasodilators; EDRF-endothelium derived relaxantfactor. Phosphodiesterase inhibitors: new opportunities for the treatment of asthma 513

with an improved side effect profile and an Table 1 Characteristics ofphosphodiesterase isozymes increased therapeutic index would be an important advance in the treatment of asthma. K,,, (pmol/l)t

Many pharmaceutical companies have attem- Family* Isozyme* cAMP cGMP Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from pted to improve the therapeutic profile of theophylline by synthesising and evaluating Il+ Ca2` /CaM stimulated 30 3 I+ Ca2 /CaM stimulated 1 2 new xanthine analogues without specifically II cGMP stimulated 50 50 focusing on the phosphodiesterase inhibitory III cGMP inhibited 0-2 0.3 activity. So far these efforts have not been IV cAMP specific 2 100 successful. V cGMP specific 150 1 An alternative approach toward developing *Nomenclature from ref 15. an "improved theophylline" has emerged tNumbers represent approximate values (see ref 14); kinetic characteristics vary substantially, depending on species, tissue, recently. This approach is based on the isolation procedure, and enzyme purity. premise that multiple, distinct forms of phos- +The subsets of PDE I are arbitrarily designated l and 1$. phodiesterases exist and that different are responsible for cyclic nucleotide hydrolysis in different tissues. Attempts are being made to increase tissue -has a much greater affinity for cGMP (Km = 3 selectivity by synthesising inhibitors that are pmol/l) than for cAMP (Km = 30 umol/l), targeted against the major phosphodiesterase whereas the second (PDE I,) has an equal isozyme or isozymes responsible for regulat- affinity (Km = 1-2 umol/l) for the two cyclic ing cAMP and cGMP hydrolysis in airway nucleotides. PDE II (cGMP stimulated phos- smooth muscle and inflammatory cells. phodiesterase) has a high Km for both cAMP Although some of these inhibitors have and cGMP, and displays positive cooper- retained the xanthine nucleus, others have a ativity with respect to substrate. The activity completely different chemical structure. It is of this enzyme against cAMP is stimulated hoped that the phosphodiesterase inhibitors manyfold by "physiological" concentrations that emerge will retain at least some of the (0 1-1-0 pmol/l) of cGMP. The isozyme with anti-inflammatory or bronchodilator activity the lowest Km for cAMP and cGMP (about (or both) of theophylline but that they will be 0-3 umol/l for both cyclic nucleotides) is PDE less likely to produce gastrointestinal, central III. In contrast to PDE II, the ability of PDE nervous system, and cardiovascular side effects III to hydrolyse cAMP is inhibited by low by virtue of their poor activity against adeno- concentrations (0-1-10 umol/l) of cGMP. sine receptors and the phosphodiesterases PDE IV is termed the cAMP specific phos-

present in gut, brain, and heart. The purpose phodiesterase because its affinity for cAMP http://thorax.bmj.com/ of this review is to explain the theoretical (Km = 2 jymol/l) is much greater than its foundation for the new generation of phos- affinity for cGMP (Km = 100 ,umol/l). The phodiesterase inhibitors in the treatment of final isozyme, PDE V, is called the cGMP asthma. specific phosphodiesterase. This enzyme hydrolyses cAMP with a Km of more than 100 Phosphodiesterase isozymes umol/l and cGMP with a Km of 1 iimol/l. Hydrolysis of the 3'-phosphoester bond on An important point is that, whereas certain cAMP or cGMP converts these second mes- isozymes possess a degree of selectivity for on September 30, 2021 by guest. Protected copyright. sengers to their inactive 5'-nucleotide cAMP or cGMP, all phosphodiesterases are metabolites (figure). This reaction is catalysed capable of hydrolysing both cyclic nucleo- by a heterogeneous group of isozymes collec- tides. Certain isozymes may also, for various tively known as phosphodiesterases. The gen- reasons, be inactive in intact tissues. Thus, in eral characteristics of phosphodiesterase the absence of additional information, it is isozymes are shown in table 1. It is important impossible to predict which isozyme or to note that the kinetic values shown are isozymes are responsible for cyclic nucleotide approximations. Actual values may vary sub- hydrolysis in a tissue containing multiple stantially, depending on the tissues, species, phosphodiesterases simply by determining isolation procedure, and enzyme purity. The which isozymes are present in tissue homo- enzymes differ nevertheless in their kinetic and genates. physical characteristics, substrate (cAMP or From the standpoint of drug discovery the cGMP) selectivities, sensitivity to endogenous synthesis of compounds that possess a sub- activators and inhibitors, susceptibility and stantial degree of selectivity for one isozyme response to phosphorylation by protein kin- by comparison with others has been critical.2' ases, tissue distribution, and subcellular local- These compounds are valuable pharmaco- isation."4 5 Subclasses of isozymes may exist logical probes with which to assess the impor- for each of the five families of phosphodies- tance of various phosphodiesterase isozymes terases shown in table 1.L' Recent evidence in regulating cyclic nucleotide content in suggests that the five families of phosphodies- intact tissues. More importantly, isozyme terase isozymes are coded for by distinct but selective phosphodiesterase inhibitors may perhaps related genes.'l20 have therapeutic advantages over non-selec- Phosphodiesterase (PDE)I, or the Ca2 /cal- tive compounds (see below) and several iso- modulin stimulated phosphodiesterase, as its zyme selective phosphodiesterase inhibitors name implies, is stimulated by Ca'+/cal- are, or were, being assessed in clinical studies modulin. At least two general forms of the for various disorders. Some examples are enzyme have been identified; the first (PDE I,) shown in table 2. 514 Torphy, Undem

Table 2 Examples of isozyme selective phosphodiesterase (PDE) inhibitors IC50 Class Drug (/smol/l)* Source Indication or action Development staget Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from 1. Calmodulin stimulated (PDE I) 15 Wyeth-Ayerst + Phase III 2. cGMP inhibited (PDE III) (BL 4162A) 0 05 Bristol-Myers Squibb Antithrombotic Phase II (OPC 3689) 0 005 Otsuka Vasodilator Discontinued (MDL 17043) 1 Marion Merrell Dow Inotropic Phase III Imazodan (CI 914) 6 Warner-Lambert Inotropic Discontinued (WIN 47203) 0-3 Sterling Winthrop Inotropic Discontinued§ (SK&F 94836) 1 SmithKline Beecham Inotropic Discontinued 3. cAMP specific (PDE IV) Denbufylline (BRL 30892) 1 SmithKline Beecham Multi-infarct Phase III Ro 20-1724 5 Hoffmann-LaRoche Research Discontinued (ZK 62711) 1 Schering AG Antidepressant Discontinued Tibenelast (LY 186655) 20 Lilly Antiasthma Phase III TVX 2706 3 Troponwerke GmbH Anti-inflammatory Discontinued 4. Mixed PDE III/IV Benzafentrine (AH 21-132) 1-3 Sandoz Antiasthma Discontinued (B 842-90) 1-3 Byk-Gulden Antiasthma Phase II 5. cGMP specific (PDE V) 1 Commercial Vasodilator Marketed MY-5445 0-5 Mitsubishi Kasei Antithrombotic Preclinical SK&F 96231 1 SmithKline Beecham Bronchodilator Preclinical (M&B 22,948) 1 Rhone-Poulenc Rorer Antiallergic Discontinued *Value represents IC,O against corresponding PDE isozyme. Each compound possesses a selectivity for the isozyme noted, although the degree of selectivity of the different compounds ranges from fivefold to several orders of magnitude. tReflects development stage in the United States only. represent a class of drugs that improve cognitive functions, such as learning, memory, reasoning, and understanding. §Development of oral form discontinued; an intravenous formulation was approved in the United States but has not yet been marketed. IC,O-concentration causing a 50% inhibition of PDE activity.

Therapeutic hypothesis the efficacy of PDE inhibitors is that their Two important hypothetical considerations ability to increase cyclic nucleotide content is underpin the proposed use of isozyme selective related to the basal activity of adenylate or phosphodiesterase inhibitors to treat asthma. guanylate cyclase in target tissues, a con- The first is that the tissue distribution of sequence of the synergism between phos- phosphodiesterase isozymes is heteroge- phodiesterase inhibitors and activators of neous,1422-25 so that the isozymes responsible adenylate cyclase (for example, beta adren- for the regulation of cyclic nucleotide content oceptor agonists) or guanylate cyclase (for differs from tissue to tissue. Secondly, it should example, nitrovasodilators). The effect of theo- http://thorax.bmj.com/ be possible to develop organ selective phos- phylline in asthma has been attributed to phodiesterase inhibitors by targeting the com- potentiation of the effects of endogenous pounds at the predominant isozyme or iso- activators ofadenylate cyclase."l' The impor- zymes in the tissue of interest,2'l 26 27 and this has tance of phosphodiesterase inhibition in the been achieved with the PDE III selective therapeutic activity of theophylline has, inhibitors.2"30 These compounds are directed however, been questioned,213" primarily against the major cAMP hydrolysing phos- because the plasma concentrations of unbound phodiesterase present in the myocardium and theophylline that elicit bronchodilatation in on September 30, 2021 by guest. Protected copyright. vasculature and thus are intended as inotropes vivo produce only a modest (10-20%) inhibi- possessing a modest degree of vasodilator tion of phosphodiesterase activity in human activity. Although a positive effect of PDE III lung extracts.'3'4 This degree of inhibition in inhibitors on long term survival in patients the presence of endogenous adenylate or with congestive heart failure has yet to be guanylate cyclase activators may, however, be shown, these compounds have a pronounced sufficient to produce a substantial increase in beneficial effect on haemodynamics.2930 In con- cyclic nucleotide content in airway smooth trast to previous agents that lacked isozyme muscle and inflammatory cells.532 This proposal selectivity (such as theophylline), the PDE III is supported by the fact that therapeutic con- inhibitors produced few, ifany, side effects that centrations of theophylline act synergistically could be ascribed to phosphodiesterase inhibi- with exogenously administered activators of tion in non-cardiovascular tissues.25 This sup- adenylate cyclase to inhibit human polymor- ports the proposal that the side effect profile of phonuclear leucocyte activation ex vivo." 35 phosphodiesterase inhibitors can be improved In contrast to the results with human poly- by synthesising compounds that possess a high morphonuclear leucocytes, most studies in degree of selectivity for the appropriate asthmatic subjects have detected only an isozyme. An improved therapeutic index may additive bronchodilator effect when a beta permit higher doses of isozyme selective adrenoceptor agonist and theophylline are inhibitors to be administered-that is, doses given together,3638 though a synergistic that produce a substantially greater inhibition interaction has been observed on occasion.'9 of the relevant isozyme than do non-selective One possible explanation for the failure to inhibitors. The clinical efficacy of isozyme detect a synergistic interaction in vivo is that selective phosphodiesterase inhibitors may theophylline produces bronchodilatation by then be greater than that of their non-selective several cellular mechanisms in addition to counterparts. phosphodiesterase inhibition. The second factor to consider with regard to Several endogenously released agents are Phosphodiesterase inhibitors: new opportunities for the treatment of asthma 515

capable of stimulating adenylate cyclase surface receptor. A distinct particulate guany- activity in inflammatory cells and airway late cyclase is stimulated by the atriopeptins smooth muscle, and could thus act to potentiate through a receptor mediated mechanism that is

the responses to phosphodiesterase inhibitors. not dependent on a guanine nucleotide binding Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from These include (1) circulating catecholamines protein. f-rom the adrenal medulla4'; (2) prostaglandin The intracellular target enzymes for cAMP E2 (PGE2) from monocytes, alveolar macro- and cGMP are cAMP dependent and cGMP phages,41 and airway epithelial cells42; (3) pros- dependent protein kinases. As cellular concen- tacyclin (PGI2) from vascular endothelial trations ofcAMP or cGMP rise the appropriate cells43; and (4) neurotransmitters from pulmon- cyclic nucleotide binds to and activates its ary adrenergic neurones and inhibitory non- corresponding protein kinase. The activated adrenergic non-cholinergic (i-NANC) protein kinases then mediate the physiological neurones (for example, vasoactive intestinal responses to cyclic nucleotides by phos- polypeptide).4 45 The role of these endogenous phorylating and thus changing the activity of adenylate cyclase activators in regulating the key substrates (such as enzymes and ion trans- responsiveness of target cells to phosphodies- port systems) concerned in the regulation of terase inhibitors is speculative. Local concen- smooth muscle tone. Several relevant sub- trations ofautacoids such as PGE2 and PGI2 are strates are likely to be phosphorylated by increased in areas of inflammation and after cAMP dependent and cGMP dependent antigenic stimulation,46 so the functional res- protein kinases in airway smooth muscle.9 10 In ponse to phosphodiesterase inhibitors may be brief, the increases in cAMP or cGMP content greater in the inflamed lungs of patients with can produce airway smooth muscle relaxation asthma than in non-inflamed tissues. Cyclo- by two general mechanisms. Firstly and per- oxygenase inhibitors and beta adrenoceptor haps most importantly, an increase in cyclic antagonists could reduce the therapeutic nucleotide concentrations leads to a decrease in activity of phosphodiesterase inhibitors, the cytosolic free Ca2' concentration by reduc- though to our knowledge neither of these ing the mobilisation of Ca2+ from intracellular possibilities has been tested. stores, inhibiting the influx of extracellular Ca2', stimulating Ca2+ efflux or increasing Ca2+ sequestration into intracellular storage sites. Airway smooth muscle Secondly, activation of the cAMP or cGMP ROLE OF CYCLIC NUCLEOTIDES protein phosphorylation cascades may inhibit The mechanisms by which cyclic nucleotides the activity of contractile proteins directly. mediate smooth muscle relaxation has been the Several of these biochemical pathways are 4749 subject of several recent reviews.' 10 An in likely to be activated simultaneously and act in http://thorax.bmj.com/ depth discussion of this topic is beyond the a coordinated fashion to reduce airway smooth scope of this review, but a brief overview of the muscle tone. proposed roles ofcAMP and cGMP in regulat- Bronchodilators that act through second ing airway smooth muscle tone is given; further messenger pathways-for example, beta ad- details can be obtained elsewhere.9 10 47 48 renoceptor agonists and phosphodiesterase As depicted schematically in the figure, inhibitors-should relax airway smooth muscle intracellular cyclic nucleotide content is regardless ofthe mediator or mediators respon- increased by increasing the rate at which cyclic sible for inducing bronchoconstriction. This nucleotides are formed or by decreasing the should be an advantage in a disease such as on September 30, 2021 by guest. Protected copyright. rate at which they are degraded. Activators of asthma in which several mediators may act in adenylate cyclase (such as beta adrenoceptor concert to elicit bronchoconstriction. agonists and prostaglandins E2 and 12) bind to cell surface receptors and, through a guanine PHOSPHODIESTERASE ISOZYMES IN AIRWAY nucleotide binding protein, increase the rate at SMOOTH MUSCLE which Mg2+-ATP is converted to cAMP. In Over the last 15 years, and especially the last contrast to the regulation of adenylate cyclase, five, several reports have described phos- agents such as the nitrovasodilators (for exam- phodiesterase isozyme profiles in airway ple, sodium nitroprusside, nitroglycerin) and smooth muscle from various species (table endothelium derived relaxant factor (probably 3).5056 A few caveats must be considered in any nitric oxide) activate soluble guanylate cyclase discussion of the results of these studies. activity directly without interacting with a cell Firstly, kinetic analyses ofenzyme activity have generally been carried out with crude or at best partially purified preparations. To our know- Table 3 Summary ofphosphodiesterase (PDE) isozymes in airway smooth muscle ledge, there have been no reports of studies in which every phosphodiesterase isozyme Species Tissue Isozymes Reference(s) present in airway smooth muscle has been Canine Trachealis I/, II, III, IV, V 51, 52 purified to homogeneity. Secondly, airway Bovine Trachealis I, II, IV*t 53, 54 smooth muscle preparations are not homogen- Guinea pig Trachealis III, IV+ 51 eous in terms of cellular content and at least Human Bronchus I, III, IV, V*§ 55 Human Trachealis II, 1,, II, III, IV, V** 56 some of the phosphodiesterase activity present in homogenates of airway smooth muscle will *Subclassification of PDE I (that is, s or 11) has not yet been determined. come from submucosal glands, capillary endo- tFraction designated PDE I may also contain PDE V. +At least two additional fractions of phosphodiesterase activity eluting from anion-exchange thelial cells, fibroblasts, neurones, mast cells, columns were not characterised. etc. Thirdly, as discussed later, the presence of §PDE II may be present but not detected. **Results based on preliminary characterisation. a phosphodiesterase isozyme in a tissue homo- 516 Torphy, Undem

genate does not necessarily indicate a Finally, a limited amount of work has been physiological role for that isozyme. Fourthly, directed toward defining the phosphodiesterase only recently have highly selective and potent isozyme profile of human airway smooth mus-

inhibitors of phosphodiesterase isozymes been cle. In an early study by Bergstrand and Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from available as pharmacological probes; many Linquist,55 at least four distinct peaks of PDE published studies on the functional responses activity were isolated from human bronchial of airway smooth muscle to phosphodiesterase tissue by anion exchange chromatography. The inhibitors have used compounds that are not kinetic characteristics of the different activities isozyme selective. Finally, virtually all studies suggest the presence of PDEs I, III, IV, and V. describing phosphodiesterase activity in airway The fact that this study was carried out with smooth muscle have concentrated on tracheal bronchial tissue, a preparation containing a smooth muscle. We do not yet know whether highly heterogenous cell population, raises the isozyme profiles differ between central and question of whether one or more of the PDEs peripheral airways. identified were present in contaminating cells Phosphodiesterase isozyme profiles have rather than in the smooth muscle. The results been studied more extensively in the canine are supported, however, by a recent prelimin- trachealis than in any other type of airway ary report on human tracheal smooth muscle,56 smooth muscle. Polson and colleagues used which appears to contain all the phosphodies- anion exchange chromatography to separate terases present in bronchial tissue as well as phosphodiesterases from the soluble fraction of PDE I,56 an enzyme that would probably not canine trachealis homogenates almost a decade have been detected by the techniques used in ago.50 Five apparently distinct peaks of phos- the earlier study. phodiesterase activity eluted from the column, though only one of these was characterised kinetically. The phosphodiesterase in this peak ACTION OF ISOZYME SELECTIVE was not stimulated by Ca +/calmodulin and PHOSPHODIESTERASE INHIBITORS ON AIRWAY hydrolysed cAMP with a Km of 0O6 ,mol/l and SMOOTH MUSCLE cGMP with a Km of 3 pmol/l. On the basis of Identification ofa PDE isozyme in a tissue does the characteristics outlined in table 1, these not provide information on the relative impor- data suggest that at least one of the isozymes in tance of the isozymes in regulating cyclic canine trachealis is PDE III. More recently, at nucleotide content in vivo. One approach to least two groups have used conventional defining the physiological role of phosphodies- chromatographic techniques to separate canine terase isozymes, and hence appropriate tracheal phosphodiesterases and have then molecular targets for new therapeutic agents, is identified the isozymes present by means of to examine the biochemical and functional kinetic analyses, isozyme specific modulators of effects of isozyme selective inhibitors in intact http://thorax.bmj.com/ enzyme activity (for example, cGMP, Ca" / tissues. Polson and colleagues determined the calmodulin), and isozyme selective phospho- potency of various xanthines as inhibitors of diesterase inhibitors.5' 52 These studies have two distinct phosphodiesterase activities shown five distinct phosphodiesterase iso- isolated from canine trachealis: a low Km cAMP zymes, including phosphodiesterases I:, II, phosphodiesterase, probably PDE III, and a III, IV, and V. There appears to be a differen- second phosphodiesterase activity, perhaps tial subcellular distribution of these isozymes representing a mixture of isozymes, which

57 on September 30, 2021 by guest. Protected copyright. in that PDE II is absent from high speed preferred cGMP as a substrate.50 When membrane fractions, whereas PDE III is en- potencies of the various xanthines as inhibitors riched in membrane fractions.52 Although five of the two phosphodiesterase activities were isozymes are present in canine trachealis, some then related to their potencies as relaxants of make a minimal contribution to total cyclic methacholine contracted trachealis strips a nucleotide phosphodiesterase activity in tissue remarkably good correlation was observed,50 57 homogenates.9 suggesting that the xanthines examined were Bovine and guinea pig tracheal phosphodies- relaxing canine trachealis by inhibiting the terase isozymes have also been isolated and activity of a specific cAMP or cGMP phos- characterised. Cytosolic fractions of bovine phodiesterase. The results of these studies, the trachealis homogenates contain phosphodies- first to describe a correlation between inhibition terases I, II, and IV.53 54 In view ofthe substan- of distinct phosphodiesterase activities and tial inhibitory potency of zaprinast against the relaxation of airway smooth muscle, cannot be enzyme described as phosphodiesterase I"35 interpreted unequivocally. One problem was perhaps, as in canine trachealis,5'2 the enzyme the use of alkylxanthines that inhibited cAMP preparation designated PDE I in these studies and cGMP phosphodiesterase activities non- also contained PDE V. PDE III was conspic- selectively. Consequently, the potency order of uously absent from the soluble fraction of the compounds for inhibition of the cAMP bovine trachealis homogenates; whether it is phosphodiesterase activity50 was identical to present in the membrane fraction of this tissue that for inhibition of the cGMP phosphodies- is not yet known. Isolation and partial charac- terase activity.57 This and the lack of informa- terisation of phosphodiesterases from guinea tion on the effect of these compounds on cyclic pig tracheal smooth muscle indicate the nucleotide content in intact trachealis strips presence of both PDE III and PDE IV.5" At makes it difficult to determine whether the least two other peaks of phosphodiesterase relaxant response of the trachealis to the phos- activity were eluted from anion exchange phodiesterase inhibitors was due to inhibition columns, but the enzyme activity in these of either the cAMP or the cGMP phosphodies- fractions were not characterised.5' terase activity, or both. Phosphodiesterase inhibitors: new opportunities for the treatment of asthma 517

More recently, the physiological role of Finally, a role for PDE V in regulating cGMP various phosphodiesterase isozymes in canine content in guinea pig airway smooth muscle is trachealis has been examined by correlating the suggested by the observation that SK&F

functional response (relaxation) of the tissue to 96231, a selective PDE V inhibitor, relaxes Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from isozyme selective phosphodiesterase inhibitors guinea pig isolated trachea in a concentration with the biochemical response (cyclic nucleo- dependent manner.65 tide content) to these compounds.58 59 The biochemical response of guinea pig Siguazodan (SK&F 94836), a highly selective trachealis to isozyme selective phosphodies- PDE III inhibitor,5260 produced a concentra- terase inhibitors has not been studied in great tion dependent relaxation of methacholine or detail. The PDE III inhibitor SK&F 94120 histamine contracted canine trachealis strips and the mixed PDE III/IV inhibitor benzafen- that was associated with an increase in cAMP trine (AH 21-132) increase cAMP content in content and an activation of cAMP dependent muscle rich guinea pig tracheal preparations, protein kinase.58 On the other hand, siguazodan though detectable increases in cAMP content had no effect on the functional or biochemical are observed only with concentrations of these (that is, cGMP accumulation) responses to compounds that produce near maximal relaxa- sodium nitroprusside,58 a guanylate cyclase tion.6' 6 Much lower concentrations of ben- activator that is thought to relax airway smooth zafentrine substantially potentiated the ability muscle via a cGMP mediated mechanism.'0 of a threshold concentration of forskolin to These results provide strong evidence to sup- elevate cAMP content in this preparation. port a major role for PDE III in regulating Benzafentrine,' but not SK&F 94120,61 also cAMP content but not cGMP content in intact increases cGMP content in guinea pig tra- canine trachealis. chealis. SK&F 96231, a PDE V inhibitor, A similar approach was taken in another increased cGMP content in guinea pig lung study aimed at defining the roles of PDE III, parenchymal strips.65 Although the strips con- IV, and V in canine trachealis.59 These tain many cell types in addition to airway experiments assessed the ability of various smooth muscle, this observation implies that isozyme selective PDE inhibitors to potentiate inhibition ofPDE V in intact tissues can lead to the functional and biochemical responses to a rise in cGMP content. isoprenaline or sodium nitroprusside. The The bronchorelaxant activity of isozyme phosphodiesterase inhibitors used were SK&F selective phosphodiesterase inhibitors in the 94120, Ro 20-1724, and zaprinast, all of which guinea pig is not limited to in vitro prepara- had been shown to selectively inhibit canine tions. Various PDE III, PDE IV, mixed PDE tracheal PDE III, PDE IV, and PDE V respec- III/IV, and PDE V inhibitors have substantial

tively.52 Isoprenaline induced relaxation and bronchodilator activity in anesthetised guinea http://thorax.bmj.com/ cAMP accumulation were potentiated by both pigs63 65 67-71 and, as in experiments in vitro, the SK&F 94120 and Ro 20-1724, but these agents bronchodilatation appears to occur regardless had no effect on the response to sodium of the spasmogen used to induce tone. nitroprusside. In contrast, zaprinast poten- The effects ofisozyme selective phosphodies- tiated sodium nitroprusside induced relaxation terase inhibitors in bovine airway smooth and cGMP accumulation, but did not poten- muscle have not been examined as extensively tiate the responses to isoprenaline. These as in guinea pig or canine airway smooth results suggest that PDE III and PDE IV muscle and there are no data correlating relax- hydrolyse cAMP in intact canine trachealis, ant responses with changes in cyclic nucleotide on September 30, 2021 by guest. Protected copyright. whereas PDE V hydrolyses cGMP.59 content. Preliminary data suggesting that Several studies have shown the potent relax- rolipram is a potent relaxant of bovine tra- ant activity of PDE III inhibitors on the guinea chealis imply an important role for PDE IV in pig isolated trachea, providing further support regulating cAMP content in this tissue.5472 On for a role for this isozyme in regulating airway the other hand, the role of PDE III in bovine smooth muscle cAMP content.6'"" Harris and trachealis is somewhat ambiguous. Although colleagues found an excellent correlation be- the PDE III inhibitors SK&F 94120 and tween the ability of a series of compounds to milrinone relax this tissue, the potency and relax guinea pig trachea and their ability to efficacy of both compounds are distinctly inhibit PDE III isolated from guinea pig low.54 72 The poor relaxant activity of PDE III trachealis.6' An analogous series ofexperiments inhibitors, coupled with the failure to detect conducted with a different set of compounds PDE III in the soluble fraction of bovine showed no relation between relaxation and trachealis homogenates,53 54 raises doubt about PDE IV inhibition but, inexplicably, relaxa- the functional role of PDE III in this tissue. tion did correlate with the ability of these The functional role of PDE V in the bovine compounds to compete with a high affinity trachea appears to be minor as zaprinast has rolipram binding site.6' The nature and func- little or no relaxant activity in this tissue.54 1 73 tion of this high affinity rolipram binding site Functional effects of isozyme selective phos- have yet to be defined,'M but it may represent a phodiesterase inhibitors on human isolated discrete subtype of PDE IV. There was pro- airway smooth muscle have not yet been repor- nounced synergism between the relaxant ted, though several groups are actively engaged activity of PDE III inhibitors and compounds in this area of research and results of these that bind to the high affinity rolipram binding studies should be available soon. site. For example, the EC50 for rolipram was One final caveat concerns the interpretation reduced nearly 25 fold in the presence of a fixed of the data discussed in this section: no phos- concentration of CI-930, a PDE III inhibitor.6" phodiesterase inhibitors possess absolute 518 Torphy, Undem

isozyme specificity. When present in high con- below) are now known to have inhibitory effects centrations all isozyme selective inhibitors will on many inflammatory cell types, including inhibit phosphodiesterases other than the ones eosinophils, neutrophils, monocytes, platelets, 76-79 against which they are targeted. Certain basophils, and mast cells.5 112 Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from isozyme selective phosphodiesterase inhibitors Although many pharmacological studies may also regulate physiological processes by have shown that increasing cAMP content mechanisms that are unrelated to phosphodies- inhibits inflammatory cell activity, there is little terase inhibition. For example, the ability of information on the specific mechanisms mediat- rolipram to inhibit agonist induced generation ing this inhibitory effect, though it appears to be of inositol phosphates in bovine trachealis may due to multiple mechanisms. As with the be unrelated to phosphodiesterase inhibition studies on smooth muscle discussed above, and cAMP accumulation.73 Thus conclusions evidence indicates that at least one of the on the role of individual phosphodiesterase mechanisms by which cAMP suppresses in- isozymes in intact tissues drawn from experi- flammatory cell activity includes an inhibitory ments with isozyme selective phosphodies- effect on stimulus induced increase in cytosolic terase inhibitors are somewhat tenuous, parti- Ca2+ .78 80-85 Because an increase in cytosolic cularly in the absence of corroborative bio- Ca2+ is a trigger for several cellular activities, chemical data. including mediator production and release, inhibiting this event would be expected to cause a general inhibition of inflammatory cell Inflammatory cells function. There is also evidence, however, that ROLE OF CYCLIC NUCLEOTIDES cAMP can inhibit mediator release in the The vast number of published papers leaves absence of gross changes in cytosolic Ca2+ little doubt that cAMP acts as an inhibitory concentrations.84 In mast cells a rise in cAMP is messenger in inflammatory cells.5" 12 In addi- much more effective in inhibiting antigen tion to inhibiting inflammatory mediator induced eicosanoid biosynthesis than in inhi- production and release, it is evident that an biting the rise in cytosolic Ca2+ or histamine increase in cellular cAMP inhibits other func- release.8486 In the case of eicosanoid biosyn- tions of inflammatory cells, including chemo- thesis, it has been suggested that activation of taxis, cytotoxicity and cell aggregation. Thus the cAMP cascade may cause direct inhibition increases in cAMP content have a generalised of an enzyme or enzymes in the biosynthetic suppressive influence on human inflammatory pathway (for example, phospholipase A2).84 cells (figure). Much less is known about the role of cGMP More than 20 years ago Lichtenstein and in the regulation of inflammatory cell function. in cellular An increase in cGMP may enhance antigen coworkers suggested that an increase http://thorax.bmj.com/ cAMP led to an inhibition of antigen induced induced histamine release from human lung histamine release from human basophils.74 mast cells.87 These studies, however, were This speculation was based on the observation carried out on lung parenchymal fragments, so that beta adrenoceptor agonists and theo- the biochemical data are difficult to interpret. phylline were effective inhibitors of antigen Neither sodium nitroprusside nor dibutyryl induced histamine release from human mixed cGMP inhibited immunological mediator leucocyte preparations, and was consistent release from human basophils88 or isolated with the observations made by Schild some 30 human lung mast cells in our studies (unpub- years earlier that epinephrine is a potent lished observations) Although the precise role on September 30, 2021 by guest. Protected copyright. inhibitor of anaphylactic histamine release of cGMP in regulating inflammatory cell func- from lung tissue.75 Various compounds that tion is uncertain, its effect appears to be modest either stimulate adenylate cyclase or inhibit by comparison with the profound inhibitory cyclic nucleotide phosphodiesterase (see effect of cAMP.

PHOSPHODIESTERASE ISOZYMES IN Table 4 Summary ofphosphodiesterase (PDE) isozymes in human inflammatory cells INFLAMMATORY CELLS The profile of phosphodiesterase isozymes in Predominant inflammatory cells has been obtained only PDE Effect of PDE recently (table 4). Chromatographic analyses Cell type isozyme(s) t inhibitors+ Reference(s) combined with kinetic and pharmacological Basophil IV lMediator release 88 characterisations show that PDE IV is the Mast cell IV iMediator release 32 major phosphodiesterase isozyme responsible Neutrophil IV lMediator release 90, 99 iRespiratory burst for catabolising cAMP in several human inflam- Eosinophil* IV ISuperoxide formation 76 matory cell types, including mast cells, baso- Monocyte IV iMediator release 77, 91, 93, 103, 104 phils, monocytes, and neutrophils. In contrast, ICytokine formation in Lymphocyte III, IV iCytotoxicity 91, 92, 98, 100, 101 PDE III is the predominant isozyme plate- lIgE formation lets and a combination of PDE III and PDE IV ICytokine formation is present in mixed lymphocyte preparations. iBlastogenesis at the of Platelet IIITAggregation 58, 89, 94, 105, 106 The next section looks identity iMediator release the predominant cAMP phosphodiesterase ac- tivity in various inflammatory cells. In virtually *So far only studies with guinea pig eosinophils have been conducted. all the cells studied so far, however, minor tOnly the major cAMP metabolising enzyme or enzymes are listed. $Represents the effect of either isozyme selective or non-selective phosphodiesterase inhibitors; activity from one or more phosphodiesterases although a few of the cited studies were conducted with non-selective compounds, the overall has been found in addition to the major pharmacological effect observed is likely to have been due to inhibition of the major phosphodiesterase isozyme present in the cell. activity. In the basophil and platelet, for exam- Phosphodiesterase inhibitors: new opportunitiesfor the treatment of asthma 519

ple, the additional phosphodiesterase is selec- Platelet tive for cGMP and is inhibited by zaprinast,88"8 The predominant phosphodiesterase isozyme suggesting that it is PDE V. On the basis of the in the human platelet is PDE III... .4. This

substrate specificity this isozyme seems unlikely isozyme has a Km for both cAMP and cGMP of Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from to have a physiological role in the catabolism of about 0-4 imol/l, is selectively inhibited by cAMP. Thus selectively inhibiting this iso- several PDE III inhibitors, including cilo- zyme is unlikely to lead to a rise in cellular stamide, milrinone, and siguazodan, and is cAMP. inhibited by cGMP (IC,, - 0-1 imol/l), a distinct characteristic of PDE III. Basophil and mast cell In the human basophil most of the cAMP ACTION OF ISOZYME SELECTIVE INHIBITORS ON hydrolysis activity is accounted for by PDE IV INFLAMMATORY CELL FUNCTION (Km = 1 6 jmol/1).88 A small portion (under In view of the individual profile of phos- 20%) of the activity is attributed to PDE III phodiesterase isozymes in inflammatory cells, (Km = 0 3 1M), on the basis ofthe observations and knowledge that a rise in cAMP has potent that the low Km isozyme was inhibited by anti-inflammatory activity, it is reasonable to cGMP and the selective PDE III inhibitor propose that isozyme selective inhibitors may SK&F 95654. The phosphodiesterase isozyme be useful as anti-inflammatory agents in the profile in human mast cell populations is not treatment of asthma. Information on the func- known. Mouse mast cells derived from mouse tional consequence of selective inhibition of bone marrow contain a single cAMP hydrolys- phosphodiesterase isozymes in human inflam- ing isozyme with kinetic and pharmacological matory cells is summarised below. Because characteristics of PDE IV.32 much of the work is recent some conclusions are drawn from data that are unpublished or published only as abstracts. Appropriate Neutrophil caution in these areas is warranted. In The cAMP PDE activity in the human neutro- general is the PDE isozyme profile in the respective cell phils eluted from DEAE anion exchange type appears to be a good predictor of the columns in a single peak, has a Km of0 7 pmol/l, pharmacology of isozyme-selective inhibitors and is sensitive to inhibition by the selective (table 4). PDE IV inhibitors rolipram and Ro 20-1724.9o These findings suggest that the phosphodies- Mast cell terase isozyme responsible for hydrolysing SK&F 95654 and zaprinast, inhibitors of PDE cAMP in the human neutrophil is predomi- III and PDE V respectively, have little or no nantly if not exclusively PDE IV.

effect on immunologically stimulated mediator http://thorax.bmj.com/ release from human isolated lung mast cells.32 Lymphocyte In contrast, rolipram and Ro 20-1724 inhibit Systematic evaluation of the PDE isozyme antigen induced mediator release from these profile in the various subtypes of lymphocytes cells, and potentiate the inhibitory effect of has not been performed to our knowledge. adenylate cyclase activators.32 The effect of Thompson and coworkers reported that a isozyme selective inhibitors on human lung mixed lymphocyte preparation (99% lym- mast cells is mimicked in the mouse bone phocytes) contains a single phosphodiesterase marrow derived mast cell. In these cells PDE isozyme with a low Km (0 40 ymol/l) for cAMP IV inhibition reduces antigen induced his- on September 30, 2021 by guest. Protected copyright. hydrolysis and a much higher Km for cGMP tamine release and leukotriene C4 production, hydrolysis.9' The elution profile and kinetic whereas zaprinast and siguazodan are without characteristics suggest that the isozyme is PDE effect.32 Interestingly, the rat peritoneal mast III. Both PDE III and PDE IV may be present cell appears to respond differently to isozyme in T lymphocytes because in preparations selective inhibitors from both the human lung enriched in T lymphocytes two high affinity mast cell and the mouse bone marrow derived cAMP selective phosphodiesterases exist, one mast cell. In this cell type submicromolar of which is inhibited by Ro 20-1724 and concentrations of zaprinast effectively inhibit another which is not.92 The different isozyme immunologically induced histamine release, profile observed in these cells versus mixed whereas rolipram has little effect.95 Whether this lymphocytes raises the possibility that different reflects a species difference or is a result of subsets of lymphocyte may contain different microenvironment dependent mast cell phosphodiesterases .92 heterogeneity is unknown. Monocyte Basophil Purified monocyte preparations isolated from Inhibiting PDE IV with rolipram or Ro 20- peripheral blood contain a single cAMP selec- 1724 effectively inhibits antigen-induced his- tive phosphodiesterase with a Km of 1-3 tamine and leukotriene C4 release from human ,mol/l19; this isozyme appears to be PDE IV. basophils.889596 Rolipram also potentiates the This is supported by a preliminary study in inhibitory effect offorskolin on mediator release which the major cAMP phosphodiesterase in from these cells.95 Neither zaprinast nor SK&F monocytes was partially purified and character- 95654 influences antigen induced mediator ised93 and shown to have the kinetic character- release from the human basophil, but SK&F istics (cAMP Km = 1-8 ymol/l) and pharmaco- 95654 potentiates the inhibitory effect of logical properties (inhibited by rolipram and rolipram. This potentiation is observed both Ro 20-1724) consistent with its being PDE IV. for inhibition of mediator release and increase 520 Torphy, Undem

in cellular cAMP.95 Thus there appears to be a terase inhibitors on monocyte function have synergistic interaction between the inhibitors not been reported. of PDE III and PDE IV in the basophil.

Combinations of PDE III and PDE IV inhibi- Platelet Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from tors have also been reported to act synergis- Consistent with its phosphodiesterase isozyme tically in guinea pig trachealis, rat ventricle, profile, platelet function is profoundly inhibited and human lymphocyte.639798 Examination of by selective PDE III inhibitors. Sub- the effects of various combinations of isozyme micromolar concentrations of several selective selective phosphodiesterase inhibitors on in- PDE III inhibitors, including anegralide, SKF flammatory cell function deserves further 94120, Ro 15-2041, and siguazodan, are effec- attention. tive inhibitors of agonist induced calcium mobilisation, aggregation, and secre- Neutrophil tion.60 891051' Inhibition of PDE IV or V with The neutrophil respiratory burst is inhibited rolipram and zaprinast respectively has little or by the PDE IV inhibitors, rolipram, and Ro 20- no effect on platelet responses.89 1724.9 The effectiveness of these inhibitors is equal to that of the relatively non-selective phosphodiesterase inhibitors theophylline and Clinical experience . The phosphodiesterase profile of THERAPEUTIC EFFECTS these cells suggests that PDE IV is the only Clinical experience with the use of isozyme phosphodiesterase isozyme regulating cAMP selective phosphodiesterase inhibitors for the content in neutrophils. Thus selective concen- treatment of asthma is limited. To our know- trations of zaprinast and various PDE III ledge, the first such compound to be examined inhibitors are without effect on the respiratory was zaprinast, an inhibitor ofthe cGMP specific burst in this cell type. Inhibition of phos- phosphodiesterase (PDE V). In an initial phodiesterase and increase in cAMP have also placebo controlled, double blind crossover trial been found to inhibit lysozyme release from oral zaprinast (10 mg) reduced exercise in- neutrophils and the synthesis of platelet duced bronchoconstriction in adult asthmatic activating factor and LTB,.7999 patients significantly,'07 whereas it failed to inhibit bronchoconstriction induced by his- Eosinophil tamine. The reason that zaprinast was active No information is available on the effect of against exercise induced bronchoconstriction isozyme selective phosphodiesterase inhibitors but not histamine induced bronchoconstriction on human eosinophil function. In a prelimin- is not obvious. On the basis ofthe proposed role ary report Dent and coworkers reported that of cGMP in airway smooth muscle, we might PDE IV selective inhibitors reduce superoxide speculate that the efficacy ofzaprinast stemmed http://thorax.bmj.com/ formation in guinea pig eosinophils.76 from bronchodilator activity. This explanation does not appear to be viable, however, in view Lymphocyte ofthe failure ofzaprinast to abrogate histamine The PDE IV inhibitor Ro 20-1724 inhibits the induced bronchoconstriction. On the other activity of human cytotoxic T lymphocytes'°° hand, zaprinast might have altered neural and interleukin-2 production by human T- reflexes or mediator release, perhaps through a lymphocytes."'0 In addition, both Ro 20-1724 mechanism unrelated to PDE inhibition. But it and CI-930, a PDE III inhibitor, cause a had no effect on exercise induced bronchocon- on September 30, 2021 by guest. Protected copyright. modest inhibition of blastogenesis in T lym- striction in a subsequent study with asthmatic phocytes.98 These agents have a much more children.'08 Thus the positive results obtained pronounced inhibitory effect on blastogenesis in the earlier study were not repeated in a when used in combination, however, suggest- different group of asthmatic patients. ing that both PDE III and PDE IV activities At least a dozen PDE III inhibitors have are important in regulating cAMP content in been evaluated for inotropic or antithrombotic these cells.9 98 Selective inhibitors of PDE IV activity in man but bronchodilator activity has may also inhibit antibody production by B been assessed only for one, enoximone.'09 Pul- lymphocytes. Ro 20-1724 inhibits IgE produc- monary mechanics were studied before and tion by peripheral blood mononuclear leuco- after an intravenous infusion of enoximone (3 cytes and most, but not all, of this activity can mg/kg) in 19 patients with decompensated be attributed to a direct inhibitory effect on the chronic obstructive pulmonary disease. Enox- B cells.'02 imone reduced pulmonary resistance and increased dynamic lung compliance in both Monocyte spontaneously breathing and artificially ven- Inhibition of monocyte phosphodiesterase and tilated patients. Although long term studies the resulant rise in cellular cAMP is associated have not been conducted in patients with with inhibition of arachidonic acid metabolism asthma, these preliminary results suggest that in zymosan treated cells." Theophylline and enoximone can reverse bronchoconstriction in other agents that increase cAMP content also man. inhibit the production of tumour necrosis Preliminary results on the activity of tiben- factor in human monocytes and rat macro- alast,"10 a PDE IV inhibitor, and benzafen- phages,'03104 probably by a cytokine specific trine,"1' a mixed PDE III/IV inhibitor, also mechanism as phosphodiesterase inhibitors have been reported. The results with tibenalast have no effect on the production ofinter-leukin- were equivocal. A single oral 150 mg dose of 1.103 Effects of isozyme-selective phosphodies- tibenalast caused a slight but non-significant Phosphodiesterase inhibitors: new opportunities for the treatment of asthma 521

increase in FEV, in 40 asthmatic subjects. Conclusions Because of the putative anti-inflammatory Understanding of the potential role of isozyme activity of PDE IV inhibitors, it would be of selective phosphodiesterase inhibitors in the interest to assess the effect of tibenalast over a treatment of asthma is far from complete. Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from longer time. Inhalation of benzafentrine (2-32 Preliminary data are beginning to provide an mg) by non-asthmatic subjects reversed metha- overall sense of the potential value of these choline induced bronchoconstriction, as compounds. Most evidence so far suggests that monitored by changes in specific airways con- PDE III inhibitors, and perhaps PDE IV and ductance, in a dose dependent manner."' A V inhibitors, possess bronchodilator activity. preliminary report indicates that intravenously In addition, inhibitors of PDE IV may have administered benzafentrine (40 and 80 mg) activity against the inflammatory aspects of produced a modest and transient broncho- asthma. Little is known about the potential dilatation, whereas the compound was in- utility of inhibitors of PDE I and PDE II. effective when administered by the oral route Attempts to broaden the therapeutic activity (9_90 mg)."2 of isozyme selective phosphodiesterase inhibitors have been made by synthesising SIDE EFFECTS hybrid compounds and at least two mixed PDE With the possible exception of the PDE III III/IV inhibitors, zardaverine and benzafen- inhibitors, clinical experience with isozyme- trine, have undergone preliminary clinical selective phosphodiesterase inhibitors is not studies. These compounds may provide an yet extensive enough to draw conclusions about optimal combination of bronchodilator and the side effect profile of the different classes of anti-inflammatory activities. Hypothetically, a compounds. The potential for certain side combination of PDE IV and PDE V inhibitors effects can be predicted based on the anticipated would have a similar therapeutic profile. activity ofdifferent isozyme selective inhibitors Although the concept of hybrid phosphodies- on various organs.97 For example, by virtue of terase inhibitors is superficially attractive, any their cardiotonic activity, PDE III inhibitors loss in isozyme selectivity is likely to lead to would be expected to produce an increase in more side effects, thus subverting one of the myocardial contractility and modest vasodila- major proposed advantages of isozyme selec- tion.'>30 Whether such effects would be tive phosphodiesterase inhibitors over non- detrimental in subjects with asthma who were selective compounds. otherwise healthy is unknown. Perhaps ofmore After having been moribund for many years, concern is the arrhythmogenic potential of interest in phosphodiesterase inhibitors as these compounds."3 If such activity is the agents for asthma has undergone a striking

result of PDE III inhibition per se, it could resurgence, particularly with respect to the http://thorax.bmj.com/ represent a major limitation to the use of these potential therapeutic advantages of isozyme agents to treat asthma. selective phosphodiesterase inhibitors. An The most obvious concern regarding the understanding of the range of activity of these PDE IV inhibitors stems from their apparent compounds should emerge over the next antidepressant activity."s"6 The anti-depres- decade as results from clinical trials become sant activity of rolipram and, presumably, available. other compounds of its class appears to be linked to inhibition of PDE IV in the brain."7 We gratefully acknowledge the help of Dr Carol Kulp for conducting literature searches and Ms Dotti Lavan for her on September 30, 2021 by guest. Protected copyright. Whether PDE IV inhibitors will produce un- careful preparation of this manuscript. acceptable central nervous system effects in individuals who do not suffer from affective disorders 1 Jenne JW. Physiology and pharmacodvnamics of the xan- remains to be answered. Gastrointes- thines. In: Jenne JW, Murphy S, eds. Drug therapy for tinal disturbances, such as pyrosis, nausea, and asthma: research and clinical practice. (Lung biology in vomiting, are frequently reported side health and disease series.) New York: Dekker, 1987: effects 297-334. associated with the use of rolipram.11"l6 2 PerssonCGA.Overviewofeffectsoftheophylline. JAllergy Rolipram has also been reported to produce a Clin Immunol 1986;78:780-7. 3 Persson CGA, Erjefilt I, Gustafason B. Xanthines- rapid and pronounced fall in plasma osmo- symptomatic or prophylactic in asthma? Agents Actions lality, probably as a result of an antidiuretic 1988;23(Suppl):S 1 37-55. effect; this resolves after seven days' treat- 4 Weinberger M. Treatment of chronic asthma with theo- phylline. ISI Atlas Sci Pharm 1988;2:53-61. ment.'8 It is not clear whether these effects are 5 Kuehl FA, Zanetti ME, Soderman DD, Miller DK, Ham due to PDE IV inhibition in general or whether EA. Cyclic AMP-dependent regulation of lipid mediators they are unique to in white cells. A unifying concept for explaining the rolipram. efficacy of theophylline in asthma. Am Rev Respir Dis Cyclic GMP is purported to have a critical 1987;136:210-3. physiological role in mediating vasodilation 6 Fredholm BB. Are methylxanthine effects due to antag- onism of endogenous adenosine? Trends Pharmacol Sci and PDE V appears to be important in regulat- 1980;1: 129-33. ing cGMP content in vascular smooth mus- 7 Poisner AM. Direct stimulant effect of on catecholamine release from the adrenal medulla. Biochem cle.471'9 120 Thus PDE V inhibitors may have Pharmacol 1973;22:469-76. cardiovascular activity. 8 Fink G, Mittelman M, Shokat B, Spitzer SA. Theo- Because of the dearth of information con- phylline-induced alterations in cellular immunity in asth- matic patients. Clin Allergy 1987;17:313-6. cerning the clinical activity of isozyme-selec- 9 Torphy TJ. Action of mediators on airway smooth muscle: tive phosphodiesterase inhibitors, most of the functional antagonism as a mechanism for bronchodilator side drugs. Agents Actions 1 988;23(suppl):S37-53. effects discussed above are speculative. 10 Torphy TJ, Hay DWP. Biochemical regulation of airway Individual agents are also likely to produce side smooth muscle tone: an overview. In: Townley RG, effects that are not related to phosphodiesterase Agrawal DK, eds. Airway smooth muscle: modulation of receptors and response. Boca Raton: CRC Press, 1990: inhibition and cannot therefore be predicted. 39-68. 522 Torphy, Undem

11 Bourne HR, Lichtenstein LM, Melmon KL, Henney CS, 40 Barnes PJ. Endogenous catecholamines and asthma. Weinstein Y, Shearer GM. Modulation of inflammation Allergy Clin Immunol 1986;77:791-5. and immunity by cyclic AMP. Science 1974;184:19-28. 41 Godard P, Chaintreuil J, Damon M, et al. Functional 12 Plaut M, Marone G, Thomas LL, Lichtenstein LM. Cyclic assessment of alveolar macrophages: comparison of cells nucleotides in immune responses and allergy. AdvCyclic from asthmatics and normal subjects. J Clin Allergy Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from Nucleotide Res 1980;12:161-72. Immunol 1982;70:88-93. 13 Church MK, Featherstone RL, Cushley MJ, Mann JS, 42 Keikauf GD, Ueki IF, Nadel JA, Widdicombe JH. Brady- Holgate ST. Relationships between adenosine, cyclic kinin stimulates chloride secretion and prostaglandin E2 nucleotides, and xanthines in asthma. J Allergy Clin release by canine tracheal epithelium. AmJ Physiol 1985; Immunol 1986;78:670-5. 248:F48-55. 14 Beavo JA. Multiple isozymes of cyclic nucleotide 43 Johnson AP. Human pulmonary endothelial cells in cul- phosphodiesterase. Adv Second Messenger Phosphoprotein ture: activities of cells from arteries and cells from veins. Res 1988;22:1-38. Clin Invest 1980;65:841-50. 15 Beavo JA, Reifsnyder DH. Primary sequence of cyclic 44 Laitinen LA, Laitinen A. Innervations of airway smooth nucleotide phosphodiesterase isozymes and the design of muscle. Am Rev Respir Dis 1987;136:S38-42. selective inhibitors. Trends Pharmac Sci 1990;1 1:150-5. 45 Palmer JBA, Barnes PJ. Neuropeptides and airway smooth 16 Colicelli J, Birchmeier C, Michaeli T, O'Neill K, Riggs M, muscle function. Am Rev Respir Dis 1987;136:S50-4. Wigler M. Isolation and characterization of a mammalian 46 Schulman ES, Adkinson NF, Jr, Newball HH. Cyclooxy- gene encoding a high-affinity cAMP phosphodiesterase. genase metabolites in human lung anaphylaxis: airway vs. Proc Natl Acad Sci USA 1989;86:3599-603. parenchyma.J Appl Physiol 1982;53:589-95. 17 Henkel-Tigges J, Davis RL. Rat homologs of the Droso- 47 Lincoln TM. Cyclic GMP and mechanisms of vasodilation. phila dunce gene code for cyclic AMP phosphodiesterase Pharmac Ther 1989;41:479-502. sensitive to rolipram and RO 20-1724. Mol Pharmacol 48 Murray KJ. Cyclic AMP and mechanisms of vasodilators. 1990;37:7-10. Pharmac Ther 1990;47:329-45. 18 Pittler SF, Baehr W, Wasmuth JJ, McConnell DF, 49 Krall JF. Receptor-mediated regulation of tension in Champagne MS, van Tuinen P, Ledbetter D, Davis RL. smooth muscle cells. In: Jenne JW, Murphy S, eds. Drug Molecular characterization of human and bovine rod therapy for asthma: research and clinical practice. New photoreceptor cGMP phosphodiesterasex-subunit and York: Dekker, 1987:97-128. chromosomal localization of the human gene. Genomics 50 Polson JB, Krzanowski JJ, Szentivanyi A. Inhibition of a 1990;6:272-83. high affinity cyclic AMP phosphodiesterase and relaxation 19 Swinnen JV, Joseph DR, Conti M. Molecular cloning of rat of canine tracheal smooth muscle. Biochem Pharmacol homologues of the Drosophila melanogaster dunce cAMP 1983;31:3403-6. phosphodiesterase: evidence for a family of genes. Proc 51 Silver PJ, Hammel LT, Perrone MH, Bentley RG, NatI Acad Sci USA 1989;86:5325-9. Bushover CR, Evans DB. Differential pharmacologic 20 Livi GP, Kmetz P, McHale MM, et al. Cloning and sensitivity of cyclic nucleotide phosphodiesterase expression of cDNA for a human low-K,,, rolipram- isozymes isolated from cardiac muscle, arterial and airway sensitive cyclic AMP phosphodiesterase. Molec CellBiol smooth muscle. Eur J Pharmacol 1988;150:85-94. 1990;10:2678-86. 52 Torphy TJ, Cieslinski LB. Characterization and selective 21 Weishaar RE, Cain MH, Bristol JA. A new generation of inhibition of cyclic nucleotide phosphodiesterase phosphodiesterase inhibitors: multiple molecular forms of isozymes in canine tracheal smooth muscle. Molec phosphodiesterase and the potential for drug selectivity. Pharmacol 1990;37:206-14. JMedChem 1985;28:537-45. 53 Elliott KRF, Berry JL, Bate AJ, Foster RW, Small RC. 22 Beavo JA, Hardman JG, Sutherland EW. Hydrolysis of The isoenzyme selectivity of AH 21-132 as an inhibitor of cyclic guanosine and adenosine 3',5'-monophosphates by cyclic nucleotide phosphodiesterase activity. J Enzyme rat and bovine tissues.J Biol Chem 1970;245:5649-55. Inhibition 1990;4:245-5 1. 2'3 Appleman MM, Avaro MA, Takemato DJ, Whitson RH. 54 Nicholson CD, Shahid M, van Amsterdam RGM, Cyclic nucleotide phosphodiesterases. In: Nathanson JA, Zaagsma J. Cyclic hucleotide phosphodiesterase (PDE) Kebalian JW, eds. Handbook of experimental pharma- isoenzymes in bovine tracheal smooth muscle, their cology. Vol 5812. Berlin: Springer, 1982:262-300. inhibition and the ability of isoenzyme inhibitors to relax http://thorax.bmj.com/ 24 Thompson WJ, Appleman MM. Characterization of cyclic precontracted preparations. Eur J Pharmacol 1990;183: nucleotide phosphodiesterases of rat tissues. J Biol Chem 1097-8. 1971;246:3145-50. 55 Bergstrand H, Lundquist B. Partial purification and charac- 25 Wells JN, Hardman JG. Cyclic nucleotide phosphodies- terization of cyclic nucleotide phosphodiesterase from terases. Adv Cyclic Nucleotide Res 1977;8:119-43. human bronchial tissue. Molec Cell Biochem 1978;121: 26 HidakaH, Tanaka T, Itoh H. Selective inhibitors of cyclic 9-15. nucleotide phosphodiesterases. Trends Pharmacol Sci 56 Cieslinski LB, Reeves ML, Torphy TJ. Cyclic nucleotide 1984;5:237-9. phosphodiesterase (PDEs) in canine and human tracheal 27 Weiss B. Differential activation and inhibition of multiple smooth muscle (TSM) [abstract]. FASEB J 1988;2: forms of cyclic nucleotide phosphodiesterase. Adv Cyclic A 1065. Nucleotide Res 1975;-5:195-21 1. 57 Polson JB, Krzanowski JJ, Szentivanyi A. Correlation 28 Colucci WS, Wright RF, Braunwald E. New positive between inhibition of a cyclic GMP phosphodiesterase on September 30, 2021 by guest. Protected copyright. inotropic agents in the treatment of congestive heart and relaxation of canine tracheal smooth muscle. Biochem failure: mechanisms of action and recent clinical develop- Pharmacol 1985;34:1875-9. ments. J 1986;314:349-58. N Engl Med 58 Murray KJ, England PJ, Hallam TJ, et al. The effects of 29 Hood WB. Controlled and uncontrolled studies of phos- siguazadon, a selective phosphodiesterase inhibitor, on phodiesterase III inhibitors in contemporary cardio- human platelet function. Br J Pharmacol 1990;99:612-6. vascular medicine. Am J Cardiol 1989;63:46-53A. 59 Torphy TJ, Burman M, Huang LBF, Tucker SS. Inhibi- 30 Wood MA, Hess ML. Review: long-term oral therapy of tion of the low K, cyclic AMP phosphodiesterase in intact congestive heart failure with phosphodiesterase trachealis by SK&F 94836: mechanical and biochemical inhibitors. Am JMed Sci 1989;297:105-13. J 31 Nielson CP, Crowley JJ, Cusack BJ, Vestal RE. responses. Pharmacol Exp Ther 1988;246:843-50. Therapeutic concentrations of theophylline and 60 Torphy TJ, Zhou H-L, Burman M, Huang LBF. Role of enprofylline potentiate catecholamine effects and inhibit cyclic nucleotide phosphodiesterase isozymes in intact leukocyte activation. J Allergy Clin Immunol 1986;78: canine trachealis. Molec Pharmacol 1991;39:376-84. 660-7. 61 Bryson SE, Rodger IW. Effects of phosphodiesterase 32 Torphy TJ, Livi GP, Bacarek JM, White JR. Chilton FH, inhibitors on normal and chemically-skinned isolated airway smooth muscle. Br J Pharmacol 1987;92:673-81. Undem BJ. Therapeutic potential of isozyme-selective 62 Rossing TH, Drazen JM. Effects of milrinone on contractile phosphodiesterase inhibitors in the treatment of asthma. responses of guinea pig trachea, lung parenchyma and Second Messenger Phosphoprotein Research (in press). J 33 Bergstrand H. Phosphodiesterase inhibition and theo- pulmonary artery. Pharmacol Exp Ther 1986;238:874-9. phylline. Eur J Respir Dis 1980;10(suppl):37-44. 63 Harris AL, Connell MJ, Ferguson EW, et al. Role of low K, cyclic AMP phosphodiesterase inhibition in tracheal 34 Polson JB, Krzanowski JJ, Goldman AL, Szentivanyi A. J Inhibition of human pulmonary phosphodiesterase relaxation and bronchodilation in the guinea pig. activity by therapeutic levels of theophylline. Clin Exp Pharmacol Exp Ther 1989;251:199-206. Pharmacol Physiol 1978;8:535-9. 64 Schneider HH, Schmiechan R, Brezinski M, Seidler J. 35 Nielson CP, Crowley JJ, Morgan ME, Vestal RE. Stereospecific binding of the antidepressant rolipram to Polymorphonuclear leukocyte inhibition by therapeutic brain protein structures. Eur J Pharmacol 1986;127: concentrations of theophylline is mediated by cyclic-3',5'- 105-15. adenosine monophosphate. Am Rev Respir Dis 1988;137: 65 Murray KJ, Eden RJ, England PJ, et al. Potential use of 25-30. selective phosphodiesterase inhibitors in the treatment of 36 Bitling B, Dahlqvist R, Garli M, Hornbland Y, Ripe E. asthma. Agents Actions (in press). 66 Small RC, Berry JL, Boyle JP, et Biochemical and Separate and combined use of terbutaline and theo- al. phylline in asthmatics. Effects related to plasma levels. Eur electrical aspects of the tracheal relaxant action of AH 21- J Respir Dis 1982;63:399-409. 132. Eur JPharmacol 1991;192:417-26. 37 Handslip PDJ, Dart AM, Davies BH. Intravenous salbu- 67 Gristwood RW, Sampford KA. Inhibition of histamine- induced bronchoconstriction by SK&F 94836, salbutamol tamol and aminophylline in asthma: a search for synergy. and theophylline [abstract]. Br J Pharmacol 1987;92: Thorax 1981;36:741-4. 631P. 38 Svedmyr K, Svedmyr N. Does theophylline potentiate 68 Osborn RR, Ferracone JD, Novak LB, Hand JM. Effects of inhaled /l, agonists? Allergy 1982;37:101-10. rolipram (RO) enantiomers on in vitro and in vivo guinea 39 Svedmyr N, Svedmyr K. In vitro and in vivo effects of J theophylline and (J-adrenostimulants in combination. Eur pig pulmonary models [abstract]. FA SEB 1990;4:A615. J Respir Dis 1980;109(suppl):83-91. 69 Bewlev JS, Chapman ID. AH 21-132, a novel relaxant of Phosphodiesterase inhibitors: new opportunities for the treatment of asthma 523

airway smooth muscle [abstract]. Br J Pharmacol 1988; cyclic AMP- and cuyclic GMP-phosphodiesterase 93:52P. inhibitors on immunological release of histamine and on 70 Beume R, Eltze M, Kilian U. Bronchodilation due to lung contraction. Br J Pharmac 1981;73:933-8. inhibition of cAMP-PDE III: preclinical results on the 96 Busse WW, Anderson VL. The granulocyte response to the pyridazinone derivative B 842-90. Arch Pharmacol 1987; phosphodiesterase inhibitor Ro 20-1724 in asthma. 335(suppl):R76. J Allergy ClGn Immunol 1981;67:70-4. 71 Ho PPK, Esterman M, Wang L, et al. LY 186655, 5,6- 97 Nicholson CD, Challiss RAJ, Shahid M. Differential Thorax: first published as 10.1136/thx.46.7.512 on 1 July 1991. Downloaded from diethoxybenzo(B)thiophene-2-carboxylic acid, sodium modulation of tissue function and therapeutic potential of salt, a new orally active bronchodilator [abstract]. selective inhibitors of cyclic nucleotide phosphodiesterase Pharmacologist 1987;29:173. isoenzymes. Trends Pharmacol Sci 1991;12:19-27. 72 Hall IP, Walker D, Hill SJ, Tattersfield AE. Effect of 98 Robicsek SA, Blanchard DK, Djeu J, Krzanowski JJ, isozyme selective phosphodiesterase inhibitors on bovine Szentivanyl A, Polson JB. Attenuation of human T-lym- tracheal smooth muscle tone. Eur J Pharmacol phocyte blastogenesis by the selective inhibition of 1990;183: 1096-7. cAMP-phosphodiesterase [abstract]. FASEB J 1990;4: 73 Hall IP, Donaldson J, Hill SJ. Modulation of carbachol- A337. induced inositol phosphate formation in bovine tracheal 99 Chilton FH, Schmidt D, Torphy T, Goldman D, Undem smooth muscle by cyclic AMP phosphodiesterase inhibi- B. cAMP inhibits platelet activating factor (PAF) bio- tors. Biochem Pharmacol 1990;39:1357-63. synthesis in human neutrophil [abstract]. FASEBJ 1989; 74 Lichtenstein LM, Margolis S. Histamine release in vitro: 3:A308. inhibition by catecholamines and methylxanthines. 100 Plaut M, Marone G, Gillespie E. The role of cyclic AMP in Science 1968;161:902-3. modulating cytotoxic T lymphocytes. II. Sequential 75 Schild H. Histamine release and anaphylactic shock in changes during culture in responsiveness of cytotoxic isolated lungs of guinea-pigs. Q J Exp Physiol 1936; lymnphocytes to cyclic AMP-active agents. J Immunol 26:165-79. 1983;131 :2945-52. 76 Dent G, Rabe K, Giembycz MA, Barnes PJ. Inhibition of 101 Averill LE, Kammer GM. Inhibition of interleukin 2 eosinophil respiratory burst activity by type IV-, but not production is mediated by a cyclic adenosine mono- type 111-selective cAMP phosphodiesterase inhibitors phosphate (cAMP)-dependent pathway (abstract]. Clin [abstract]. Am Rev Respir Dis 1990;141:A878. Res 1985;33:839A. 77 Godfrey RW, Manzi RM, Gennaro DE, Hoffstein ST. 102 Cooper KD, Kang K, Chan SC, Hanifin JM. Phosphodies- Phospholipid and arachidonic acid metabolism in terase inhibition by Ro 20-1724 reduces hyper-IgE syn- zymosan-stimulated human monocytes: modulation by thesis by atopic dermatitis cells in vitro. J Invest Dermatol cAMP. J Cell Physiol 1987;131:384-92. 1985;84:477-82. 78 Takenawa T, Ishitoya J, Nagai Y. Inhibitory effect of 103 Lee JC, Votta B, Dalton BJ, Griswold DE, Bender PE, prostaglandin E5, forskolin, and dibutyryl cAMP on Hanna N. Inhibition ofhuman monocyte IL-I production arachidonic acid release and inositol phospholipid by SK&F 86002. IntlJ Immunother 1990;6:1-12. metabolism in guinea pig neutrophils. J Biol Chem 1986; 104 Renz H, Gong J-H, Schmidt A, Nain A, Gemsa D. Release 261:1092-8. of tumor necrosis factor-x from macrophages. Enhance- 79 Zurier RB, Weissmann G, Hoffstein S, Kammerman S, Tai ment and suppression are dose-dependently regulated by HH. Mechanisms of lysosomal enzyme release from prostaglandin E, and cyclic nucleotides. J Immunol 1988; human leukocytes. II. Effects ofcAMP and cGMP auton- 141:2388-93. omic agonists and agents which effect microtubule func- 105 Muggli R, Tschopp TB, Mittelholzer E, Baumgartner HR. tion. J Clin Invest 1974;53:297-309. 7-Bromo-1 ,5-dihydro-3,6-dimethylimidazo(2,1 -b)quin- 80 Foreman JC, Hallett MB, Mongar JL. The relationship azolin-2(3H)-one (Ro 15-2041), a potent antithrombotic between histamine secretion and 45-calcium uptake by agent that selectively inhibits platelet cyclic AMP-phos- mast cells. J Physiol (London) 1977;271:193-214. phodiesterase. J Pharmacol Exp Ther 1985;235:212-9. 81 Nielson C. f,-Adrenergic modulation of the polymorpho- 106 Seiler S Arnold AJ, Grove RI, Fifer CA, Kelly SL, Stanton nuclear leukocyte respiratory burst is dependent upon the HC. Effects of anagrelide on platelet cAMP levels, cAMP- mechanism of cell activation. J Immunol 1987;139:2392-7. dependent protein kinase and thrombin-induced Ca++ 82 Rasmussen H, Good DBP. Relationships between calcium fluxes. J Pharmacol Exp 1987;243:767-74. and cyclic nucleotides in cell activation. Physiol Rev 107 Rudd RM, Gellert AR, Studdy PR, Geddes DM. Inhibi- 1977;57:421-509. tion of exercise-induced asthma by an orally absorbed http://thorax.bmj.com/ 83 Undem Bj, Graziano FM, Buckner CK. Effects of isopro- mast cell stabiliser (M&B 22,948). Br J Dis Chest 1983; terenol on histamine release induced from monodispersed 77:78-86. guinea pig lung cells by different secretagogues. J Phar- 108 Reiser J, Yeang Y, Warner JO. The effect of zaprinast (M&B macol Exp Ther 1985;233:617-22. 22,948, an orally absorbed mast cell stabilizer) on exercise- 84 Undem BJ, Torphy TJ, Goldman D, Chilton FH. Inhibi- induced asthma in children. Br J Dis Chest 1986;80: tion by adenosine 3':5'-monophosphate of eicosanoid and 157-63. platelet-activating factor biosynthesis in the mouse PT- 18 109 Leeman M, Lejeune P, Meot C, Naeije R. Reduction in mast cell. J Biol Chem 1987;265:6750-8. pulmonary hypertension and in airway resistance by 85 White JR, Ishizaka T, Ishizaka K, Sha'afi RK. Direct enoximone (MDL 17,043) in decompensated COPD. demonstration of increased intracellular concentration of Chest 1987;91:662-6.

free calcium as measured by quin-2 in stimulated rat 110 Israel E, Mathur PN, Tashkin D, Drazen JM. LY 186655 on September 30, 2021 by guest. Protected copyright. peritoneal mast cell. Proc Natl Acad Sci USA 1984; prevents bronchospasm in asthma of moderate severity 81:3978-82. [abstract]. Chest 1988;94:71S. 86 Undem BJ, Peachell PT, Lichtenstein LM. Isoproterenol- 111 Foster RW, Rakshi K. Bronchodilator potency, effective- induced inhibition ofimmunoglobulin E-mediated release ness and time course of inhaled nebulised AH 21-132 in of histamine and arachidonic acid metabolites from the normal human subjects [abstract]. Br J Pharmacol 1990; human lung mast cell. J Pharmacol Exp Ther 1988;247: 99: 193P. 209-17. 112 Small RC, Foster RW, Berry JL, Chapman ID, Elliott 87 Kaliner M, Orange RP, Austen KF. Immunological release KRF. The bronchodilator action of AH 21-132. Agents of histamine and slow reacting substance of anaphylaxis Actions (in press). from human lung. IV. Enhancement by cholinergic and 113 Naccarelli GV, Goldstein RA. Electrophysiology of phos- alpha adrenergic stimulation. J Exp Med 1972;136: phodiesterase inhibitors. Am J Cardiol 1989;63:35-40A. 556-67. 114 Horowski R, Sastre-Y-Hernandez M. Clinical effects of the 88 Peachell PT, Undem BJ, Schleimer RP, Lichtenstein LM, neurotropic selective cAMP phosphodiesterase inhibitor Torphy TJ. Action of isozyme-selective phosphodies- rolipram in depressed patients: global evaluation of the terase (PDE) inhibitors on human basophils [abstract]. preliminary reports. Curr Ther Res 1985;38:23-9. FASEB J 1990;4:A639. 115 Bobon D, Breulet M, Gerard-Bandenhove MA, et al. Is 89 Simpson AWM, Reeves ML, Rink TJ. Effects of SK&F phosphodiesterase inhibition a new mechanism of 94120, an inhibitor of cyclic nucelotide phosphodiesterase antidepressant action? A double blind double-dummy type III, on human platelets. Biochem Pharmacol study between rolipram and desipramine in hospitalized 1988;37:231 5-20. major and/or endogenous depressives. Eur Arch Psychiatr 90 Neilson CP, Vestal RE, Sturm RJ, Heaslip R. Effects of Neurol Sci 1988;238:2-6. selective phosphodiesterase inhibitors on the poly- 116 Eckmann F, Fichte K, Meya U, Sastre-Y-Hernandez M. morphonuclear leukocyte respiratory burst. J Allery Clin Rolipram in major depression: results of a double-blind Immunol 1990;86:801-8. comparative study with amitriptyline. Curr Ther Res 91 Thompson WJ, Ross CP, Pledger WJ, Strada SJ, Banner 1988;43:291-5. RL, Hersh EM. Cyclic adenosine 3':5'-monophosphate 117 Wachtel H. Potential antidepressant activity of rolipram phosphodiesterase. Distinct forms in human lymphocytes and other selective cyclic adenosine 3',5'-monophosphate and monocytes. J Biol Chem 197f 251:4922-9. phosphodiesterase inhibitors. Neuropharmacology 1983; 92 Robicsek SA, Krzanowski JJ, Szentivanyl A, Polson JB. 22:267-72. High pressure liquid chromatography of cyclic nucleotide 1 18 Sturgess I, Searle GF. The acute effect ofthe phosphodies- phosphodiesterase from purified human T-lymphocytes. terase inhibitor rolipram on plasma osmolality. J Clin Biochem Biophys Res Commun 1989;163:554-60. Pharmacol 1990;29:369-70. 93 White JR, Torphy TJ, Christensen SB IV, Lee JA, Mong 119 Lugnier C, Schoeffter P, LeBec A, Strouthou E, Stoclet JC. S. Partial purification of the rolipram sensitive phos- Selective inhibition of cyclic nucleotide phosphodies- phodiesterase from human monocytes [abstract]. FASEB terases ofhuman, bovine and rat aorta. Biochem Pharmacol J 1990;4:A1987. 1986;35:1743-51. 94 Macphee CH, Harrison SA, Beavo JA. Immunological 120 Souness JE, Brazdil R, Diocee BK, Jordan R. Role of identification of the major platelet low-K,, cAMP phos- selective cyclic GMP phosphodiesterase inhibition in the phodiesterase: probable target for antithrombic agents. myorelaxant actions of M&B 22,948, MY-5445, vinpoce- Proc Natl Acad Sci, USA 1986;83:6660-3. tine and 1-methyl-3-isobutyl-8-(methylamino) xanthine. 95 Frossard N, Landry Y, Pauli G, Ruckstuhl M. Effects of Br J Pharmacol 1989;98:725-35.