Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Asthma is a complex multifactorial disease characterized by reversible airway obstruction, airway inflammation, and nonspecific airway hyperreactivity (Mayer and Wills-Karp, 1999; Bertrand, 2000). Chronic obstructive pulmonary diseases (COPDs), on the other hand, are characterized by mostly irreversible airway obstruction due to chronic bronchitis and emphysema (Hay, 2000). Inflammation of the airways is believed to be central to the airways dysfunction in asthma and COPD (O'Shaughnessy et al., 1997; Roche, 1998). In these conditions, the airway wall is infiltrated by a variety of inflammatory cells, including mast cells, macrophages, T lymphocytes, eosinophils, and neutrophils. These cells release a host of mediators, including cytokines, chemokines, and bronchospastic agents that act in concert with neurotransmitters such as acetylcholine and neurokinins from pulmonary nerves to produce bronchospasm, pulmonary edema, mucus hypersecretion, and other features of asthma and COPD. Eosinophilia is the dominant feature of lung inflammation in asthma, whereas COPD is marked by an intense pulmonary neutrophilia. Although bronchodilators such as -agonists and anticholinergics are widely used for symptomatic relief, glucocorticoids are the only drugs currently available that effectively treat inflammation in asthma but not in COPD (Bertrand, 2000; Hay, 2000). One group of potential therapies for chronic pulmonary conditions is inhibitors of type 4 cAMP-specific phosphodiesterase (PDE4), of which theophylline, a nonspecific PDE inhibitor currently available for the treatment of asthma and COPD, is considered as prototypic (Barnette and Underwood, 2000).

There are at least 11 PDE enzyme families that degrade cAMP and/or cGMP (Torphy, 1998; Giembycz, 2000). In recent years, PDE4 has been widely pursued as a target to develop selective inhibitors with the hope of reducing the adverse effects associated with nonselective inhibitors such as theophylline. PDE4 is viewed as an exciting anti-inflammatory target for several reasons: 1) leukocyte functions are suppressed by cAMP, 2) PDE4 is the predominant isoform in inflammatory and immune cells, and 3) inhibitors of PDE4 negatively regulate the functions of almost all proinflammatory and immune cells and exert widespread anti-inflammatory activities in animal models of asthma. In recent years, selective PDE4 inhibitors have entered clinical trials, but most have failed due primarily to dose-limiting emesis and gastrointestinal disturbances (Martin, 2001). Cilomilast and roflumilast represent a newer generation of PDE4 inhibitors (Barnette et al., 1998; Underwood et al., 1998; Torphy et al., 1999; Bundschuh et al., 2001; Hatzelmann and Schudt, 2001) and are now in advanced stages of clinical development, showing promising efficacy in allergic rhinitis, asthma, and COPD in phase II trials (Compton et al., 1999, 2001; Schmidt et al., 2001; Timmer et al., 2002).

In this article, we describe SCH 351591 (Fig. 1) as a novel, selective, potent PDE4 inhibitor. Oral efficacy of this compound was evaluated in several animal models of asthma, including inflammatory cell recruitment into the airways and airway hyperreactivity (two cardinal features of asthma) in allergic guinea pigs and cynomolgus monkeys, and hyperventilation-induced bronchospasm (a model of exercised-induced asthma) in guinea pigs. The emetic potential of SCH 351591 was evaluated in ferrets, a widely used model of emesis. We also present limited data on SCH 365351 (Fig. 1), the major, active metabolite of SCH 351591 found in mice, rats, and monkeys, which may contribute to the efficacy of SCH 351591 in these species. We conclude from these studies that SCH 351591 exhibits a biological profile predictive of its utility in pulmonary conditions such as asthma and COPD.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Structures of SCH 351591 and SCH 365351. SCH 365351, a deoxygenated metabolite of SCH 351591, was found in variable extents in mice, rats, and monkeys.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. SCH 351591, SCH 365351, and cilomilast were synthesized at Celltech Chiroscience Ltd. (Cambridge, UK). Salbutamol was from Schering Plough (Kenilworth, NJ). [³H]cAMP, [³H]cGMP, [³H]rolipram, scintillation proximity assay (SPA) beads, and Ficoll-Paque were from Amersham Biosciences UK, Ltd. (Little Chalfont, Buckinghamshire, UK). Rolipram was from RBI/Sigma (St. Louis, MO). Ascaris suum extract was from Greer Laboratories (Lenoir, North Carolina). Polyclonal human elevated IgE sera and affinity-purified polyclonal sheep anti-human IgE antibody were from The Binding Site (San Diego, CA). Heat-killed Bordatella pertussis was from Connaught Laboratories (North York, ON, Canada). PDE enzymes were obtained as described below. All other reagents were purchased from standard laboratory supply vendors including Sigma Chemical (St. Louis, MO) and Fisher Scientific (Springfield, NJ).

PDE Enzyme Assays. The PDE activity was determined radiometrically as described previously (Wang et al., 1997) by SPA. The assay mixture contained 50 mM Tris, pH 7.5, 8.3 mM MgCl₂, 1.7 mM EGTA, various concentrations of inhibitor, and an aliquot of the enzyme solution in a final volume of 100 µl. After preincubation for 5 min at 30°C, the reaction was started by the addition of substrate (cAMP or cGMP). After incubation for an additional 30 min, the reaction was terminated by the addition of 900 µg of yttrium silicate SPA beads, and the vials were counted for radioactivity. Sufficient enzyme was added to achieve 15 to 20% substrate breakdown.

Rolipram Binding Assay. The rolipram binding assay was performed radiometrically in 96-well MAFC NOB filter plates (Millipore Corporation, Bedford, MA). The assay mixture contained 20 mM Tris-HCl, pH 7.5, 2 mM MgCl₂, 0.1 mM dithiothreitol, 100 µg of rat brain (excluding the cerebellum) membrane proteins, various concentrations of compound, and 0.5% (v/v) DMSO. Nonspecific binding was determined in the presence of 5 µM rolipram. The assay was started by the addition of 4 nM [³H]rolipram (10,000 cpm). After incubation at 22°C for 1 h, the reaction mixture was filtered, washed with 3 × 200 µl of ice-cold saline, the filters dried, 100 µl of scintillant added, and the plates left to stand for 1 h before counting for radioactivity.

Tumor Necrosis Factor- (TNF) Production by Peripheral Blood Mononuclear Cells (PBMCs). Human PBMC were isolated from buffy coats by centrifugation on a density gradient of Ficoll-Paque. PBMCs were harvested, washed three times, resuspended at 2 × 10⁶/ml in RPMI 1640 medium containing 2% fetal bovine serum, and 250-µl aliquots plated in 48-well tissue culture plates. Inhibitors were dissolved in DMSO, diluted in RPMI 1640 medium, and 100 µl added to duplicate wells at a range of concentrations (final DMSO concentration of 0.5%). Cells were stimulated with lipopolysaccharide (LPS) at a final concentration of 100 ng/ml and incubated for 18 h at 37°C in an atmosphere of 5% CO₂ and 95% air. Cells were pelleted by centrifugation, and TNF was measured in the supernatant by ELISA (R & D Systems, Minneapolis, MN).

Interleukin (IL)-5 Production by Human Peripheral Blood Leukocytes. Inhibitors were dissolved in DMSO, diluted in RPMI 1640 medium, and 100 µl added to duplicate wells of a 48-well plate at a range of concentrations (final DMSO concentration of 0.5%). Buffy coat was diluted 2-fold with RPMI 1640 medium, and 200 µl was added to each well. Cells were stimulated by the addition of 100 µl of 10 µg/ml phytohemagglutinin to a final concentration of 2.5 µg/ml and 100 µl of 500 nM phorbol-12-myristate-13-acetate solution to a final concentration of 100 nM. The plate was then incubated for 48 h at 37°C in an atmosphere of 5% CO₂ and 95% air. Cells were pelleted, and IL-5 in the supernatant was measured by ELISA (R & D Systems).

IL-12 Production by Human PBMCs. Human PBMCs were isolated from buffy coats by centrifugation on a density gradient of Ficoll-Paque. PBMCs were harvested, washed three times, resuspended at 3 × 10⁶/ml in serum-free RPMI 1640 medium, and 400-µl aliquots plated in 48-well tissue culture plates. Inhibitors were dissolved in DMSO, diluted in RPMI 1640 medium, and 100 µl added to duplicate wells at a range of concentrations (final DMSO concentration of 1%). Cells were stimulated with Pansorbin (Staphylococcus aureus suspension) at a final concentration of 0.075% and incubated for 18 h at 37°C in an atmosphere of 5% CO₂ and 95% air. Cells were pelleted by centrifugation, and IL-12 in the supernatant was measured by ELISA (R & D Systems).

Bronchodilator Activity in Isolated Guinea Pig Trachea. Cervical tracheal ring segments (5.0 mm in length) isolated from the Charles River Hartley strain of male guinea pigs (740-813 g) were equilibrated at 1.0 g of initial tension in Krebs' buffer (37°C) in the presence of indomethacin (2 µM) and contracted with histamine (10 µM). When the contraction stabilized (10-15 min), a cumulative concentration response with test compound (half-log concentration increments at 10-min intervals) was performed. The -adrenoreceptor agonist salbutamol (10 µM) was added at the end of each experiment to obtain a maximal reversal of the histamine-induced contraction. SCH 351591 and cilomilast stock solutions in DMSO were diluted with the assay buffer. Test compound-induced relaxations were normalized as percentage of maximum reversal.

Antiallergic Activity in Isolated Human Bronchus. Donor lung tissue was procured from five males and three females, 16 to 56 years old, by the International Institute for the Advancement of Medicine (Scranton, PA) and the Anatomic Gift Foundation (Woodbine, GA). Macroscopically normal lung tissue was procured after cancer resection from a 67-year-old female by the Cooperative Human Tissue Network (Eastern Division, University of Pennsylvania Medical Center, Philadelphia, PA). Hospital medications did not include corticosteroids, theophylline, or long-acting -adrenergic agents. Lung tissue was received in physiological media on wet ice 20 to 48 h after removal. Bronchus segments (4-10-mm internal diameter) were isolated, prepared as 5-mm-wide epithelium-denuded transverse muscle strips, and used on the day of arrival.

Induction of Cytokines in Corynebacterium parvum-Primed Mice. For priming with C. parvum, BDF1 mice were injected intravenously with 0.5 mg of heat-killed cells of ATCC strain 11827. One week after priming, mice were challenged i.v. with LPS (20 µg/mouse). Serum cytokine levels were quantified by cytokine-specific ELISAs. The postchallenge levels of TNF, IL-10, and IL-12 were measured in blood samples drawn at 1.5 h. Mice were dosed orally with either vehicle or SCH 351591 2 h before the LPS challenge.

LPS-Induced Pulmonary Inflammation in Rats. Male Sprague-Dawley rats (250-300 g) were anesthetized by inhalation of isoflurane (flow rate 1 ml/min; supplemented with O₂). Using a Penn-Centry microspray needle, 0.1 ml of a 100-µg/ml LPS solution in saline was injected into the trachea. Animals not challenged with the LPS solution received 0.1 ml of saline. Animals were placed on a heat pad until they recovered from anesthesia. Afterward, they were returned to their cages and allowed food and water ad libitum. All animals survived these manipulations and no additional interventions were required to ensure their survival. Animals fasted overnight were orally dosed with either cilomilast, SCH 351591, or vehicle (0.4% methylcellulose) 2 h before the LPS challenge.

Acute Allergic Bronchospasm in Guinea Pigs. Male Hartley guinea pigs were sensitized with an intraperitoneal injection (0.5 ml) of a saline suspension containing 100 mg/ml alum and 100 µg/ml ovalbumin. Additionally, each animal was primed with an intraperitoneal injection (0.3 ml) of heat-killed B. pertussis (20 optical units/ml). Animals were returned to their cages and allowed food and water ad libitum. After 27 days, animals were ready for use and were fasted overnight before study.

Guinea Pig Model of Allergic Airway Hyperreactivity and Pulmonary Inflammation. Male Hartley guinea pigs were sensitized exactly as described above. Animals fasted overnight before the study were exposed to two aerosol challenges, separated by 6 h, of either saline or 0.3% ovalbumin for 10 min each. The aerosol was generated by an ultrasonic nebulizer (model Ultra Neb99; DeVilbiss). To prevent anaphylactic bronchospasm, 30 min before the first antigen challenge, animals received the H₁ antagonist pyrilamine (10 mg/kg i.p.). Animals were returned to their cages and allowed food and water ad libitum.

Hyperventilation-Induced Bronchoconstriction in Guinea Pigs. Studies were performed on male Hartley guinea pigs ranging in weight from 400 to 600 g. The animals were fasted overnight but given water ad libitum. Anesthesia was induced by intraperitoneal injection of 50 mg/kg sodium pentobarbital. Animals were prepared with tracheal, jugular venous, and esophageal catheters and were mechanically ventilated throughout the experiment with a rodent ventilator (Harvard Apparatus, Holliston, MA). The ventilation setting used for eupneic respiration was 1.25 ml/100 g at a frequency of 50 breaths/min.

Pulmonary Changes in Allergic Monkeys. Twelve naturally allergic male monkeys (mean body weight 7.3 kg) were assigned to the study. On day 1 of the experiment, each animal was anesthetized with 10 mg/kg i.m ketamine, and anesthesia was maintained by continuous intravenous infusion of 0.05 to 0.15 mg/kg/min propofol. Animals were intubated with a cuffed endotracheal tube and intermittent positive pressure ventilation started with 100% oxygen. Blood pressure, body temperature, and arterial oxygen saturation were monitored.

Ferret Emesis Assay. Male albino and fitch ferrets (0.9-1.3 kg; Eastwoods Directory Services, Godalming, Surrey, UK) were housed in groups of five per cage with free access to food and water. Compounds dissolved in syrup BP 1:1 in water were administered orally. The animals were then transferred to individual observation cages and observed continuously for a 4-h period. The behavior was recorded by videocamera, and the tapes were subsequently played back to assess emesis. Emesis was defined as rhythmic abdominal contractions that were either associated with the expulsion of the gastrointestinal contents (i.e., vomiting) or were not associated with the expulsion of the gastrointestinal contents (retching). Data are expressed as the number of animals that responded of the total number tested per dose. A dose response was constructed for each compound and at the maximum nonemetic dose, blood samples were drawn at various time points through the left jugular vein cannulae for the determination of compound concentration in the blood.

Quantification of Compound Concentration in Blood. Blood samples drawn from monkeys, guinea pigs, and ferrets were spun down, and the plasma was harvested and stored at 70°C. Aliquots of plasma samples (40 µl) were transferred into minivials, and 100 µl of acetonitrile containing 0.4 ng/µl of an internal standard was added. After vortexing and centrifugation, the supernatant was transferred to a high-performance liquid chromatography microvial. Aliquots (30 µl) of the supernatant were injected into a TSQ 7000 LC-APCI/MS/MS system equipped with an APCI source (Thermo Finnigan, San Jose, CA). The liquid chromatographic system included a 600 S controller, a 616 pump, and a 717 plus autosampler (Waters, Milford, MA). Chromatographic separation was achieved with a reverse phase liquid chromatography column (Luna 3 µm, phenyl-hexyl, 50 × 4.6 mm; Phenomentex, Torrance, CA) using an acetonitrile/water gradient. Solvent A consisted of 20:80 acetonitrile/water, with 0.6 ml of glacial acetic acid and 0.6 ml of 90% formic acid per liter of solvent. Solvent B consisted of 100% acetonitrile with 0.6 ml of glacial acetic acid and 0.6 ml of 90% formic acid. SCH 351591 (MH⁺ = 432) and SCH 365351 (MH⁺ = 416) were quantified using selected reaction monitoring; monitoring the product ions of m/z = 254 for both target compounds. The internal standard with MH⁺ = 398 was monitored by measuring the product ion of m/z = 378. Argon gas at 2.0 millitorr was used for collision-activated dissociation of the precursor ions. Dwell time for each precursor-product ion transition was 0.3 s. Standard curve samples containing both target compounds were run in duplicate with the same sets. The method was found to be linear from 5 to 5000 ng/ml. The limit of quantification was 5 ng/ml for both compounds.

Statistics. For in vivo studies in guinea pigs, rats, and mice, comparison between groups were performed using analysis of variance, and post hoc differences were assessed using Fisher's protected least significant difference. This analysis was performed using StatView for Macintosh. Data from the monkey studies were analyzed by Student's paired t test.

Animal Handling. All studies using animals were done in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the Animal Welfare Act in a program approved by the American Association of the Accreditation of Laboratory Animals Care. Protocols used in these studies were approved by the Animal Care and Use Committee of Schering-Plough Research Institute.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Inhibition of PDE4 and Rolipram-Binding Activity. Hydrolysis of [³H]cAMP to [³H]AMP by PDE4 from human monocytic U937 cells was monitored by scintillation proximity assay. The effect of PDE4 inhibitors on high-affinity rolipram binding activity was assessed by their ability to compete with binding of [³H]rolipram to rat brain membranes in a filtration binding assay. SCH 351591 (Fig. 1, see structure) inhibited both PDE4 activity and [³H]rolipram binding in a concentration-dependent manner with IC₅₀ values of 58 and 153 nM, respectively (Table 1). SCH 365351 (Fig. 1, see structure), the only significant circulating metabolite of SCH 351591, also inhibited both PDE4 activity (IC₅₀ = 20 nM) and [³H]rolipram binding (IC₅₀ = 75 nM) with potencies greater than those of SCH 351591 (Table 1). Kinetic analyses revealed that inhibitions by these compounds of PDE4 and rolipram binding were reversible and noncompetitive (data not shown). Cilomilast inhibited PDE4 activity competitively with an IC₅₀ of 86 nM (Table 1). Cilomilast was more potent than either SCH 351591 or SCH 365351 at inhibiting high-affinity [³H]rolipram binding.

Selectivity. SCH 351591, SCH 365351, and cilomilast were tested for inhibition of PDE1, 2, 3, 5, and 7 at concentrations of 10 to 20 µM. No significant inhibition was found (data not shown). Using cloned human PDE4 subtypes, SCH 351591 and SCH 365351 were found to inhibit all four subtypes (A, B, C, and D) equally well (data not shown), whereas cilomilast showed, as reported previously (Torphy et al., 1997), a 5- to 20-fold selectivity for D subtype over the others.

In Vitro Functional Activities. SCH 351591 and SCH 365351, like cilomilast, inhibited LPS-induced TNF production by human PBMCs in a dose-dependent manner (Table 2). SCH 351591 (IC₅₀ = 59 nM) compared favorably with cilomilast (IC₅₀ = 106 nM) but was less potent than SCH 365351 (IC₅₀ = 12 nM). These compounds also inhibited IL-5 and IL-12 production by human blood leukocytes in a concentration-dependent manner (Table 2).

View this table: [in this window] [in a new window]   TABLE 2 Inhibition of LPS-induced cytokine synthesis in blood leukocytes by SCH 351591, SCH 365351, and cilomilast Cytokines released by LPS-treated human PBMCs (TNF and IL-12) or buffy coat cells (IL-5) were measured by ELISA and concentrations that caused 50% inhibition (IC50) were determined. Values are displayed as geometric means with 95% confidence limits in parentheses and number of experiments in brackets.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Inhibition by SCH 351591 and aminophylline of anti-IgE antibody-induced contraction of isolated human bronchus. Contraction of muscle strips passively sensitized with human IgE sera was induced by anti-human IgE antibody. Contraction in the presence of vehicle, 1 µM SCH 351591, or 1 µM aminophylline was normalized to the 30 µM carbachol reference contraction measured before tissue sensitization, and the values are means ± S.E.M. (n = 8-9 patient determinations). Contraction, as represented by 90-min AUC estimates, was significantly inhibited by SCH 351591 (p < 0.05, analysis of variance with Dunnett's post hoc analysis).

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Reversal of histamine-induced contractions of isolated guinea pig tracheal rings. When the contraction of tracheal rings to histamine (10 µM) stabilized, a cumulative concentration response with each test compound was performed. Salbutamol (10 µM) was added at the end of each experiment to obtain the maximal reversal. Values are means ± S.E.M. of tissue responses normalized to the maximum reversal (n = 4-8 determinations).

Inhibition of LPS-Induced Cytokine Production of Cytokine in C. parvum-Primed Mice. Mice previously primed with C. parvum produced TNF, IL-10, and IL-12 after an intravenous administration of LPS as measured by increases in serum levels of these cytokines. Production of TNF and IL-12 was inhibited by SCH 351591 in a dose-dependent manner with significant inhibition occurring at 2 mg/kg p.o. for TNF and at 10 mg/kg p.o. for IL-12 (Table 3). In contrast, IL-10 production was enhanced by SCH 351591 treatment at all doses tested (Table 3), a finding consistent with the published observation that PDE4 inhibition enhanced the production of IL-10 in LPS-stimulated murine macrophages (Kambayashi et al., 1995).

Inhibition of LPS-Induced Lung Neutrophilia in Rats. Intratracheal administration of LPS to rats caused lung inflammation characterized by the appearance of neutrophils in the BAL fluid (Fig. 4). SCH 351591 at 3 mg/kg given orally 2 h before the LPS challenge inhibited neutrophil influx by 60%. No inhibition was seen at 0.3 mg/kg. In this model, cilomilast inhibited lung neutrophilia by 70% at 10 mg/kg p.o., whereas 32% inhibition seen at 3 mg/kg was not statistically significant.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Inhibition by SCH 351591 and cilomilast of LPS-induced lung neutrophilia in rats. Sprague-Dawley male rats (5/goup) were dosed orally with indicated doses of either SCH 351591 or cilomilast for 2 h before intratracheal administration of 0.1 ml of a 100 µg/ml LPS solution in saline. Eighteen hours after the LPS challenge, animals were sacrificed and the airways flushed. The BAL fluid was analyzed for total cells and neutrophils. , p < 0.5 compared with LPS control.

Effects on Acute Bronchospasm, Airway Hyperreactivity, and Lung Inflammation in Allergic Guinea Pigs. Actively sensitized guinea pigs exposed to aerosolized ovalbumin developed an acute bronchospastic response. This response was significantly inhibited (71%) by theophylline (a nonselective PDE inhibitor) at a single oral dose of 100 mg/kg, given 2 h before the challenge (data not shown). When given orally, SCH 351591 inhibited (42%) bronchospasm significantly at 3 mg/kg (Table 4). However, in the same experiment, 21% inhibition observed at a higher dose (10 mg/kg) was not statistically significant (data not shown), suggesting that SCH 351591 exhibited only a partial efficacy in this model of acute bronchospasm.

Inhibition of Hyperventilation-Induced Bronchospasm (HIB) in Guinea Pigs. Given orally before the hyperventilation challenge, SCH 351591 dose dependently inhibited HIB with an MED of 0.3 mg/kg (Fig. 5A). SCH 351591 was about 10-times more potent than cilomilast (MED = 3 mg/kg p.o.). The nonselective PDE inhibitor aminophylline (30 mg/kg p.o.) caused an average inhibition of 43%, but this was not statistically significant (data not shown). Maximum inhibition attained by either SCH 351591 or cilomilast was 50 to 60%.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Modulation by various inhibitors of HIB in guinea pigs. HIB was induced by transiently increasing the respiratory rates from 50 to 185 breaths/min. Compounds at indicated doses were given either orally 2 h before the challenge to ascertain inhibition of bronchospasm (A) or intravenously 2 min after the challenge to ascertain reversal of bronchospasm (B). Bronchospasm was measured as pulmonary resistance (RL) and the percentage of inhibition and the percentage of reversal of bronchospasm were calculated. Values are means ± S.E.M. (n = 6-12 animals/group). , p < 0.05 compared with vehicle control.

Inhibition of Lung Inflammation in Allergic Monkeys. When challenged with aerosolized A. suum extract, naturally allergic monkeys developed pulmonary inflammation characterized by an increased presence of total cells, eosinophils, and neutrophils in the bronchoalveolar lavage fluid collected 24 h after the antigen challenge (Table 7). When animals were given SCH 351591 at a single oral dose of 3 mg/kg 2 h before the antigen challenge, the influx of eosinophils in the bronchoalveolar lavage fluid was blocked by 80%. The influx of neutrophils was also inhibited. An acute bronchospasm induced by A. suum was unaffected by SCH 351591 (Table 7). The effect on bronchial hyperreactivity could not be ascertained because of large variability among animals (data not shown).

Induction of Emesis in Ferrets. When dosed orally with SCH 351591, one of eight ferrets retched at 8 mg/kg (Table 8). No emesis was seen at 5 mg/kg despite appreciable plasma levels (C_max = 3.5 µg/ml; AUC = 26.7 µg · h/ml). The maximal no-effect dose for SCH 365351 was 6 mg/kg. Cilomilast was emetic at 3 mg/kg, but not at 1 mg/kg. At 1 mg/kg, the C_max and AUC values for cilomilast were 2.4 µg/ml and 18.3 µg · h/ml, respectively.

Therapeutic Ratio for SCH 351591. Studies in ferrets demonstrate that substantial plasma exposure of SCH 351591 and SCH 365351 can be achieved without emesis (see above). In the guinea pig efficacy model, SCH 351591 was active against hyperventilation-induced bronchospasm at an oral dose of 0.3 mg/kg with C_max and AUC values of 0.24 µg/ml and 1.7 µg · h/ml, respectively (Table 9). In this model, the efficacious dose of cilomilast against hyperventilation-induced bronchospasm was 3 mg/kg (C_max = 0.5 µg/ml; AUC = 4.9 µg · h/ml). Comparison of the plasma levels of SCH 351591 at the guinea pig efficacy dose with those at the maximal no-effect dose of 5 mg/kg in ferrets (C_max = 3.5 µg/ml; AUC = 26.7 µg · h/ml) gives a therapeutic ratio of 16 for SCH 351591 (Table 9). A similar calculation gives a therapeutic ratio of 4 for cilomilast.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Intensive effort over the last decade to develop PDE4 inhibitors for the treatment of lung inflammatory conditions such as asthma and COPD has yielded a number of potent and selective PDE4 inhibitors (Martin, 2001). Two of these inhibitors (GlaxoSmithKline's cilomilast and Byk Gulden's roflumilast) are currently in phase III, and several others are in various stages of clinical development. In this report, we describe SCH 351591 and its metabolite SCH 365351 as novel PDE4 inhibitors with in vitro potencies comparable with cilomilast. Published studies show that roflumilast is over 100-fold more potent than cilomilast at inhibiting PDE4 enzyme and in vitro TNF production (Hatzelmann and Schudt, 2001). SCH 351591 and SCH 365351 did not discriminate between the four PDE4 (A, B, C, and D) subtypes, but they were highly selective versus other PDE isozymes. Importantly, in several animal models of asthma, SCH 351591 showed good efficacy (oral ED₅₀ of 0.3-1 mg/kg) against lung inflammation, bronchial hyperreactivity, and hyperventilation-induced bronchospasm. In these assays, SCH 351591 was 10- to 30-fold more potent than cilomilast. In guinea pig and rat models of lung function and inflammation, orally administered roflumilast was shown to be 30- to 300-fold more potent than cilomilast (Bundschuh et al., 2001), suggesting that the in vivo profile of roflumilast is likely to be superior to SCH 351591.

Compared by doses and by plasma levels of parent compound required for efficacy in guinea pigs and emesis in ferrets, SCH 351591 was 3- to 4-fold less emetic than cilomilast (Table 9). Several factors may contribute to the reduced emetic activity of SCH 351591. PDE4 exists in two unique conformations (Torphy et al., 1992, 1999; Torphy, 1998). One conformer binds rolipram with high affinity, and the occupation of these sites is associated with emesis and gastrointestinal disturbances. The other conformer binds rolipram much less avidly, and inhibition of this low-affinity form is correlated with anti-inflammatory actions. SCH 351591 and SCH 365351 bind to this high-affinity conformer with reduced avidity compared with rolipram. This reduced avidity of cilomilast for high-affinity rolipram sites (about 20-fold less than that of rolipram itself) has been offered as an explanation as to why cilomilast is less emetic than rolipram (Barnette et al., 1998; Torphy et al., 1999). Recent studies using gene-disrupted mice suggest that PDE4D is more involved in emesis than PDE4B (A. Robichaud, personal communication). Notably, cilomilast inhibits PDE4D with a 5- to 10-fold greater potency than it does PDE4B, whereas SCH 351591 does not distinguish between the two subtypes. Furthermore, our unpublished data showed that cilomilast (3 mg/kg p.o.) blocked gastrointestinal mobility in rats more drastically (96%) than SCH 351591 (48% at 6 mg/kg p.o.). Thus, the improved emetic profile of SCH 351591 may be the result of several factors, including equipotency against PDE4 subtypes, reduced avidity for high-affinity rolipram binding sites, and reduced potential for gastrointestinal disturbance.

SCH 351591 exhibited a poor bronchodilatory activity in isolated guinea pig trachea (Fig. 3). However, it was more effective at inhibiting allergen-induced contraction of passively sensitized human bronchus (Fig. 2). These data are consistent with the previously suggested view that the ability of PDE4 inhibitors to suppress antigen-induced contraction is due to the inhibition of mast cell degranulation rather than to direct bronchodilation (Underwood et al., 1993, 1998). Under in vivo as well as in vitro conditions, allergen-induced contraction of human airways results from the release of mediators such as histamine and cysteinyl leukotrienes (Heaslip et al., 1992). Lung mast cells contain PDE3 and PDE4 (Giembycz, 2000), and complete inhibition of antigen-driven mediator release from these cells and of contraction of passively sensitized human airways occurs only when both PDE3 and PDE4 are inhibited (Giembycz, 2000; Schmidt et al., 2000). It is important to note that SCH 351591 (3 mg/kg p.o.) did not block allergen-induced acute bronchospasm in our allergic monkeys (Table 7). It is possible that higher doses of the compound might be needed to block mast cell mediator release in these monkeys. The finding that SCH 351591 effectively blocked lung inflammation without affecting acute bronchospasm suggests that SCH 351591 is a more potent inhibitor of inflammation than of mast cell mediator release. Reversal of hyperventilation-induced bronchospasm in guinea pigs by intravenously administered SCH 351591 (Fig. 5B) suggests that, in this model where neuropeptides released from the nerve endings are the major bronchospastic mediators, SCH 351591 could offer a significant bronchodilatory activity.

In recent clinical studies, orally administered cilomilast and roflumilast have been shown to improve lung function (forced expiratory volume in 1 s) significantly in patients with asthma and COPD (Compton et al., 1999, 2001; Timmer et al., 2002). This observation demonstrates that PDE4 inhibitors are capable of providing a clinically relevant level of bronchodilation in patients with asthma and COPD, despite the fact these compounds offer mild bronchodilatory activity in animal models. This beneficial effect on airflow may be attributed to the ability of cilomilast to attenuate bronchial hyperreactivity (Table 6) and hyperventilation-induced bronchospasm (Fig. 5A) and to amplify the cAMP-elevating effects of circulating catecholamines (Underwood et al., 1996). SCH 351591 was more potent than cilomilast against both bronchial hyperreactivity (Tables 4 and 6) and hyperventilation-induced bronchospasm (Fig. 5A) in our guinea pig models. Thus, SCH 351591 should compare favorably with cilomilast in its ability to improve forced expiratory volume in 1 s in patients with asthma and COPD.

It is believed that eosinophilic inflammation mediated at least in part by the cytokine network is a major contributor to the characteristic nonspecific bronchial hyperreactivity in asthma (Venge, 1990). Like other PDE4 inhibitors (Torphy, 1998), SCH 351591 exerted marked inhibitory effects on the accumulation in the lungs of neutrophils and eosinophils in several species, including rats (Table 3), guinea pigs (Table 4), and monkeys (Table 7) in response to diverse stimuli (ovalbumin, LPS, and A. suum). In addition, SCH 351591 inhibited the production of proinflammatory cytokine TNF, while enhancing the production of anti-inflammatory cytokine IL-10 (Tables 2 and 3). Thus, modulation of the cytokine network and inhibition of the eosinophilic inflammation might have contributed to the effectiveness of SCH 351591 against airway hyperreactivity in allergic guinea pigs. Of note, however, in this regard is the fact that cilomilast inhibited bronchial hyperreactivity at doses where eosinophil infiltration remained unaffected (Table 6), suggesting that factors other than inflammation might be involved. One such factor may well be the inhibition of the release of inflammatory neuropeptides such as substance P from the sensory nerve endings. This is suggested by our observation that hyperventilation-induced bronchospasm, a response believed to be mediated by the release of neuropeptides (Ray et al., 1989), was effectively attenuated by SCH 351591 and cilomilast (Fig. 5A). PDE4 inhibitors such as rolipram inhibit the excitatory nonadrenergic and noncholinergic neurotransmission (Undem et al., 1994) and augment the nonadrenergic and noncholinergic relaxation of bronchus (Fernandes et al., 1994), further supporting the potential impact of PDE4 inhibitors on neuronal control of pulmonary function. It is also possible that PDE 4 inhibitors affect aspects of airway remodeling such smooth muscle hypertrophy and goblet cell metaplasia, thereby contributing to their beneficial effect on airway hyperreactivity.

SCH 351591, like other PDE4 inhibitors, affected the trafficking of neutrophils into the lungs (Tables 3 and 7). Although neutrophils are generally recognized as part of the inflammatory process in COPD but not in asthma, recent data demonstrate pronounced neutrophilia in the lungs of asthma patients who are refractory to current treatments or suffer from severe exacerbations (Fahy et al., 1995; Wenzel et al., 1999; Norzila et al., 2000; Gibson et al., 2001). PDE4 inhibitors might prove beneficial in these patients.

In summary, we have identified SCH 351591 as a potent, selective, orally active PDE4 inhibitor. SCH 351591 blocked pulmonary inflammation and bronchial hyperreactivity in animal models of asthma and COPD with excellent potencies. This pharmacological profile of SCH 351591 coupled with its mild bronchodilatory activity indicates its potential utility as an effective oral therapy for asthma and COPD.