AZD9668: Pharmacological Characterization of a Novel Oral Inhibitor of Neutrophil Elastase

ABSTRACT

N-{[5-(methanesulfonyl)pyridin-2-yl]methyl}-6-methyl-5-(1- methyl-1H-pyrazol-5-yl)-2-oxo-1-[3-(trifluoromethyl)phenyl]-1,2- dihydropyridine-3-carboxamide (AZD9668) is a novel, oral inhibi- tor of neutrophil elastase (NE), an enzyme implicated in the signs, symptoms, and disease progression in NE-driven respiratory dis- eases such as bronchiectasis and chronic obstructive pulmonary disease via its role in the inflammatory process, mucus overpro- duction, and lung tissue damage. In vitro and in vivo experiments were done to evaluate the binding kinetics, potency, and selec- tivity of AZD9668, its effects in whole-blood and cell-based as- says, and its efficacy in models of lung inflammation and damage. In contrast to earlier NE inhibitors, the interaction between AZD9668 and NE was rapidly reversible. AZD9668 was also highly selective for NE over other neutrophil-derived serine proteases. In cell-based assays, AZD9668 inhibited plasma NE activity in zy- mosan-stimulated whole blood. In isolated human polymorpho-nuclear cells, AZD9668 inhibited NE activity on the surface of stimulated cells and in the supernatant of primed, stimulated cells. AZD9668 showed good crossover potency to NE from other spe- cies. Oral administration of AZD9668 to mice or rats prevented human NE-induced lung injury, measured by lung hemorrhage, and an increase in matrix protein degradation products in bron- choalveolar lavage (BAL) fluid. In an acute smoke model, AZD9668 reduced the inflammatory response to cigarette smoke as indicated by a reduction in BAL neutrophils and interleukin-1β. Finally, AZD9668 prevented airspace enlargement and small airway wall remodeling in guinea pigs in response to chronic tobacco smoke exposure whether dosed therapeutically or prophylacti- cally. In summary, AZD9668 has the potential to reduce lung inflammation and the associated structural and functional changes in human diseases.

Introduction

Neutrophils are major cellular mediators of inflammation and play an important role in a number of chronic inflam- matory lung diseases, including chronic obstructive pulmo- nary disease (COPD), bronchiectasis, acute lung injury, and acute respiratory distress syndrome. For example, in COPD, the percentage of airways with neutrophils in bronchial tissues relates to the severity of airflow obstruction measured by forced expiratory volume in 1 s (Hogg et al., 2004). Fur- thermore, sputum neutrophil numbers are correlated with peripheral airway dysfunction measured by high-resolution computed tomography (O’Donnell et al., 2004).

Neutrophil elastase (NE), an enzyme stored in the azuro- philic granules of neutrophils, is an aggressive and cytotoxic 29-kDa serine protease (Sinha et al., 1987). The high intra- cellular concentration of NE (5 mM) is contained by its tight subcellular compartmentalization. Extracellular NE activity is regulated by endogenous protease inhibitors including α1- antitrypsin (α1-AT), secretory leukoprotease inhibitor, and α2-macroglobulin (Rubin, 1996). When the extracellular free enzyme concentration exceeds the buffering capacity of en- dogenous inhibitors, NE becomes implicated in the signs, symptoms, and disease progression in inflammatory lung disorders via its role in the inflammatory process (Bergin et al., 2008), mucus overproduction (Caldwell et al., 2005), and lung tissue damage (Wright et al., 2002).

ABBREVIATIONS: COPD, chronic obstructive pulmonary disease; α1-AT, α1-antitrypsin; BAL, bronchoalveolar lavage; catG, cathepsin G; fMLP, fMet-Leu- Phe; IL, interleukin; LPS, lipopolysaccharide; NE, neutrophil elastase; Pr-3, proteinase-3; CF, cystic fibrosis; ANOVA, analysis of variance; AZD9668, N-{[5- (methanesulfonyl)pyridin-2-yl]methyl}-6-methyl-5-(1-methyl-1H-pyrazol-5-yl)-2-oxo-1-[3-(trifluoromethyl)phenyl]-1,2-dihydropyridine-3-carboxamide; ONO6818, N-[2-[5-(tert-butyl)-1,3,4-oxadiazol-2-yl]-(1RS)-1-(methylethyl)-2-oxoethyl]-2-(5-amino-6-oxo-2-phenyl-6H-pyrimidin-1-yl) acetamide; ONO5046, N-{2-[({4-[(2,2-dimethylpropanoyl)oxy]phenyl}sulfonyl)amino]benzoyl}glycine; MR889, 2-(2-thiophencarboxythio)-N-[dihydro-2(3H)-thiophenone-3-yl]- propionamide; ZD0892, N-(4-methoxybenzoyl)-L-valyl-N-[3,3,3-trifluoro-(1S)-(1-methylethyl)-2-oxopropyl]-L-prolinamide.

A number of clinical observations indicate that targeting NE might be beneficial for treating some inflammatory lung diseases. In COPD, the high prevalence of emphysema in patients with α1-AT deficiency, particularly in smokers, is believed to be caused by the unchecked action of serine pro- teases on lung tissue (Laurell and Eriksson, 1963). Lung lavage studies reveal that these patients have an increased number of neutrophils, increased levels of plasma and uri- nary desmosine (a biomarker of elastin tissue breakdown) correlated to carbon monoxide diffusion capacity (Stolk et al., 2005), and a greater NE burden in the lower airways (Mor- rison et al., 1987). There is also a positive correlation be- tween the local distribution of NE in contact with alveolar interstitial elastin and the presence of emphysematous changes in COPD lung tissue (Damiano et al., 1986). In bronchoalveolar lavage (BAL) fluid from patients with COPD, increased NE and reduced endogenous antiprotease levels are correlated with emphysema severity, with high levels of NE expressed in patients with rapidly declining lung function (Betsuyaku et al., 1996).

In addition to COPD, there is accumulating evidence that NE may be important in the pathophysiology of other chronic inflammatory lung diseases, including bronchiectasis and cystic fibrosis (CF). In bronchiectasis, a disorder character- ized primarily by neutrophilic inflammation, varying degrees of fixed airway obstruction, and acute infective exacerba- tions, high levels of NE activity have been reported in BAL (Chan et al., 2003). In CF, lung destruction is caused by obstruction of the airways due to dehydrated, thickened se- cretions, resultant endobronchial infection and an exagger- ated inflammatory response leading to the development of bronchiectasis. The inflammatory response in the airways in CF is characterized by high amounts of interleukin (IL)-8, which attracts neutrophils to the lung, with a resulting ex- cessive release of NE (Weldon et al., 2009).

In animal models, genetic deficiency or pharmacological in- tervention with small-molecule or physiological inhibitors of NE affords significant protection (approximately 60% reduction) against the proinflammatory and emphysematous effects of chronic cigarette smoke exposure (Wright et al., 2002). In addi- tion, a broader role for NE modulating mucus hypersecretion and mucociliary clearance has been reported in several in vitro and in vivo pharmacological studies (e.g., Voynow et al., 1999). With the exception of α1-AT protein agents, only one other NE inhibitor has received clinical registration. Sivelestat is currently used in Korea and Japan for the treatment of acute respiratory distress syndrome-related respiratory failure (Hayakawa et al., 2010). However, the use of sivelestat is limited by its poor pharmacokinetics, necessitating admin- istration by infusion, and the risks of organ toxicity im- posed by its irreversible interaction with the target enzyme (Kawabata et al., 1991).

N-{[5-(methanesulfonyl)pyridin-2-yl]methyl}-6-methyl-5-(1- methyl-1H-pyrazol-5-yl)-2-oxo-1-[3-(trifluoromethyl)phenyl]-1,2- dihydropyridine-3-carboxamide (AZD9668) (alvelestat; http:// www.who.int/medicines/publications/druginformation/issues/PL- 104.pdf) (Fig. 1), is a novel, orally bioavailable NE inhibitor, with the potential to be useful in neutrophil-driven inflammatory lung diseases. In this article, we report the preclinical pharmacology of AZD9668 from studies used to support the drug’s clinical development.

Fig. 1. Chemical structure of AZD9668 (alvelestat) (http://www.who.int/ medicines/publications/druginformation/issues/PL-104.pdf).

Materials and Methods

Materials. Human NE and human cathepsin G (catG), purified from human sputum, were obtained from Merck (Darmstadt, Ger- many). Human NE used in the kinetic binding studies and protei- nase-3 (Pr-3) were obtained from Elastin Products (Owensville, MO). Dog and guinea pig NE (purified from dog neutrophils) and recom- binant rodent NE were prepared in house.

Coupling reagents N-ethyl(2-dimethylaminopropyl)-carbodiimide, N-hydroxysuccinimide, 1 M ethanolamine, HBS-P buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 0.005% p20), and CM5 BIAcore sensor chips all were obtained from GE Healthcare Biosciences AB (Uppsala, Sweden). Dimethyl sulfoxide was obtained from Sigma- Aldrich (Stockholm, Sweden).

Porcine pancreatic elastase, chymotrypsin, trypsin, cytochalasin B, fMet-Leu-Phe (fMLP), lipopolysaccharide (LPS) (LPS-Escherichia coli serotype 0-111B4), and Tris-HCl were obtained from Sigma-Aldrich. The substrates for NE (human and other species) and porcine pancre- atic elastase (MeOSuc-Ala-Ala-Pro-Val 7-amido-4-methylcoumarin), trypsin (N-α-benzoyl-L-arginine 7-amido-4-methylcoumarin), and chy- motrypsin (Suc-Ala-Ala-Pro-Phe 7-amido-4-methylcoumarin) were ob- tained from VWR International (Stockholm, Sweden).

The Pr-3 substrate was obtained from Elastin Products. The catG sub- strate (Abz-Glu-Pro-Phe-Trp-Glu-Asp-Glu-EDDn), N-[2-[5-(tert-butyl)-1,
3,4-oxadiazol-2-yl]-(1RS)-1-(methylethyl)-2-oxoethyl]-2-(5-amino-6-oxo-2- phenyl-6H-pyrimidin-1-yl)acetamide (ONO6818), sivelestat [N-{2-[({4- [(2,2-dimethylpropanoyl)oxy]phenyl}sulfonyl)amino]benzoyl}glycine; ONO5046], roflumilast, dexamethasone, salbutamol, and tiotropium were synthesized in house (Department of Medicinal Chemistry, AstraZeneca R&D, Lund, Sweden).

Zymosan A from Saccharomyces cerevisiae was from Fluka (Buchs, Switzerland). Opsonized zymosan was prepared by incubating sterile (boiled) zymosan with pooled fresh human serum (50 mg zymosan/ml serum) at 37°C for 20 min. The opsonized zymosan was washed in phosphate-buffered saline and stored frozen at a concentration of 30 mg/ml.

Binding Kinetics of AZD9668. The binding kinetics between AZD9668 and human NE were analyzed using a BIAcore T100 in- strument (GE Healthcare Biosciences AB) and a direct binding as- say. NE (100 µg/ml in 10 mM acetate buffer, pH 4.5), preincubated for 10 min with AZD9668 (1 µM) to maintain active site availability, was immobilized to a CM5 sensor chip (GE Healthcare Biosciences AB) surface by amine coupling. An activated and deactivated CM5 chip surface using amine coupling was used as a control surface.

After equilibration with running buffer (0.1 M Tris-HCl, pH 7.4, containing 0.5 M NaCl buffer, pH 7.4, with 1% dimethyl sulfoxide, AZD9668 was injected over the immobilized enzyme at a flow rate of 50 µl/min, and the association rate was determined. After 1 min, running buffer was applied to the surface, and the dissociation rate was determined over 5 min. The rate of complex (AB) formation between AZD9668 (A) and the immobilized NE (B) during the sample injection was given by: d[AB]/dt = kon[A][B] — koff[AB].

The interaction data were evaluated using T100 Evaluation soft- ware and the on (kon) and off (koff) rates, and KD (koff/kon) values were determined using global fit.Potency, Selectivity, and Species Crossover of AZD9668. The potency and selectivity of AZD9668 were determined by mea- suring the cleavage of peptide substrates to products by a range of serine proteases: human NE, Pr-3, catG, pancreatic elastase, tryp- sin, and chymotrypsin. Species crossover was assessed in the same way using NE from mouse, rat, guinea pig, and dog. Substrate concentrations were chosen to be at or below their calculated Km values (i.e., the concentration of substrate that permits half-maximal rate of reaction). The final substrate concentrations were as follows: MeOSuc-Ala-Ala-Pro-Val 7-amido-4-methylcoumarin (100 µM); Boc- Ala-Ala-Nva-SBzl (60 µM); N-α-benzoyl-L-arginine 7-amido-4-meth- ylcoumarin (150 µM); Suc-Ala-Ala-Pro-Phe 7-amido-4-methylcouma- rin (200 µM); and Abz-Glu-Pro-Phe-Trp-Glu-Asp-Glu-EDDn (8 µM). Enzymes were preincubated with inhibitors for 15 min before the addition of substrate, and the amount of product was measured 90 min later (60 min for porcine pancreatic elastase) after the addition of a low-pH stopping buffer (acetate buffer, pH 4.3 containing 0.20 M sodium monochloroacetate, 0.06 M sodium acetate, and 0.14 M acetic acid).

Potency (IC50) was defined as the molar concentration of AZD9668 required to inhibit protease activity (i.e., generation of product) by 50%. pIC50 was calculated as the —log10 of the IC50. The Cheng- Prusoff equation (Cheng and Prusoff, 1973) was then used to calcu- late Ki (i.e., the enzyme inhibitor dissociation constant) using prede- termined values for Km (data not shown): Ki = IC50/(1 + [S]/Km).

The selectivity of AZD9668 was compared with two reference compounds, ONO6818 and ONO5046 (sivelestat). A broader analysis of AZD9668 selectivity was made by an external company (MDS Pharma, King of Prussia, PA) in more than 300 ligand-binding, enzyme, and ion-channel assays using a single concentration of AZD9668 (10 µM). Activity in these assays was flagged by ≥50% inhibition. Full IC50 determinations were then performed in assays
where activity was observed.

Potency of AZD9668 in Whole-Blood and Cell-Based Assays. The activity of AZD9668 was measured in human whole-blood and cell-based assays (cell-associated and explosive NE assays) and a model of epithelial cell injury.In the whole-blood assay, citrate anticoagulated human whole blood was incubated with AZD9668 for 15 min before the addition of opsonized zymosan (final concentration 0.75 mg/ml). NE activity was measured in cell-free plasma after the addition of NE substrate (71 µM final concentration in 0.1 M Tris-HCl, pH 7.4 containing 0.5 M NaCl) and incubation for 60 min at room temperature.

In the cell-associated NE assay, polymorphonuclear cells (>90% neutrophils) isolated by centrifugation of citrate anticoagulated blood through Polymorphprep (Axis-Shield, Oslo, Norway), were re- suspended at a concentration of 3 × 106/ml in Hank’s balanced salt solution and then incubated sequentially with LPS (1 µg/ml) and fMLP (10 nM) to maximize the expression of cell-associated NE, as described by Owen et al. (1995). After extensive washing in phos- phate-buffered saline to remove soluble enzyme, cell-surface NE activity was measured in the presence or absence of AZD9668 (15- min preincubation) by the addition of NE substrate (final concentra- tion 71 µM) for 60 min.

In the explosive NE assay (Yurewicz and Zimmerman, 1977), polymorphprep-purified polymorphonuclear cells (0.75 × 106/ml) were preincubated with cytochalasin B (2.5 µg/ml) for 15 min at 37°C and then substrate (66 µM final concentration), inhibitor, and 30 nM fMLP were added simultaneously. NE activity was then measured after 3 min as described above. Percentage of inhibition was calcu- lated by comparing NE activity with and without AZD9668. Potency was expressed as pIC50 as described above.

Activity of AZD9668 in In Vivo Assays. Mice (18 –20 g) were kept at 5 to 10 animals per type III macrolon wire top cage (Tecni- plast, Milan, Italy) in a room with a 12-h light/dark cycle, at 20 to 22°C and 50 to 60% humidity. Mice were fed with Lactamin R70 pellets (Lantmannen, Stockholm, Sweden) and tap water ad libitum. Rats (180 –220 g) were kept at four to six animals per type IV macrolone cage (Tecniplast) with 12-h light/dark cycle at 22°C with 50 to 60% relative humidity and fed with Lactamin R70 pellets (Lantmannen) and given tap water ad libitum. All animals were allowed to acclimatize to the standard housing conditions for 1 week before use. Health monitoring was done, without positive findings, according to Federation of European Laboratory Animal Science Associations guidelines.

Guinea pigs (Hartley) were purchased from Charles River Canada (Montreal, QC, Canada) and housed with 12-h light/dark cycle at 22°C. They were fed standard guinea pig chow, supplemented with hay bales, and water ad libitum. All procedures were approved by the University of British Columbia Animal Care Committee.

The in vivo efficacy of AZD9668 was determined in three acute and one chronic animal models. All studies were carried out in accor- dance with standards established by the Council of Europe, Astra- Zeneca Global R&D Standards for animal care, European Union Directive 86/609, and Swedish legislation. The studies were ap- proved by the regional Ethics Board, Malmoe/Lund, Sweden.

In a model of acute lung injury (based on a model developed by Tremblay et al., 2002), human NE administered as a bolus into the trachea of anesthetized (isoflurane; Forene; Abbott, Solna, Sweden) mice causes hemorrhage over a period of hours and the accumulation of red blood cells in BAL fluid. Human NE (250 U/ml and 1 ml/kg body weight), dissolved in 9 mg/ml NaCl, was given to female C57BL/ 6JBomTac mice (Taconic Europe, Ry, Denmark) intratracheally 1 h after oral administration of AZD9668 in drug vehicle (0.5% hydroxy- propyl methylcellulose in citrate buffer). Nondrug-treated control and NE-treated animals were administered either 9 mg/ml NaCl or human NE as appropriate 1 h after administration of drug vehicle. Four hours later, the mice were sacrificed by an overdose of pento- barbital and subjected to BAL. The BAL fluid was then centrifuged and the cell pellet was resuspended in 1 ml of deionized water to lyse the red blood cells. Hemorrhage was defined as the concentration of hemoglobin in BAL cell lysate and calculated by determining the absorbance at 412 nm and extrapolating the values from a hemoglo- bin reference curve.

In a similar model in rats (female HanTac: WH; Taconic Europe), human NE (250 U/ml and 1 ml/kg body weight) was administered 1 h after oral dosing with AZD9668 in drug vehicle (Miglyol 812; Sasol Germany GmbH, Hamburg, Germany). Nondrug-treated control and NE-treated animals were administered either 9 mg/ml NaCl or hu- man NE as appropriate 1 h after administration of drug vehicle. BAL samples were taken from different animal groups at 2 and 4 h for the measurement of hydroxyproline and desmosine, respectively. Hy- droxyproline is a marker for collagen breakdown, and desmosine is a marker of elastin tissue breakdown.

The concentration of total desmosines (sum of desmosine and isodesmosine) in BAL fluid was measured by a method based on liquid chromatography combined with tandem mass spectrometry. Quantification was made from a standard curve in the range 0.1 to 10 nM. At the lower limit of quantification (0.1 nM), the precision of the assay was 5.6% and at the upper limit (10 nM) it was 8.4%.

The hydroxyproline concentration in BAL fluid was measured by a method based on liquid chromatography-tandem mass spectrometry. Quantification was made from a standard curve in the range 0.33 to 13.3 µM. At the lower end of the standard curve, the precision of the assay was 2.0% and at the upper end it was 2.5%.

In an acute smoke model (van der Vaart et al., 2004), mice were allowed to acclimatize to the standard housing conditions for approx- imately 1 week. In the second week, all mice were placed in the automated smoke exposure box system (ProMech Lab AB, Malmo¨, Sweden) for increased periods of time, starting with 10 min on day 1 increasing to 50 min on day 3 to acclimatize them to the new conditions.

Female BALB/cJBomTac mice (Taconic Europe) were exposed al- ternately to cigarette smoke (12 Kentucky Research cigarettes 1R3F; University of Kentucky, Louisville, KY) and fresh air for 50 min twice daily for 4 days in a whole-body box exposure system with positions randomized. AZD9668 in drug vehicle (20% glucose and 0.1% polysorbate in water) was administered orally 45 min before each smoke exposure. Nondrug-treated air and smoked control ani- mals were exposed to drug vehicle only 45 min before exposure to fresh air or smoke as appropriate. Animals were sacrificed (as de- scribed above) 16 h after the last smoke exposure and subjected to BAL. Total and differential BAL cell counts were performed. IL-1β in undiluted BAL samples was analyzed using enzyme-linked immunosorbent assay (R&D Systems Europe Ltd, Abingdon, Ox- fordshire, UK).

In a chronic smoke model that reproduces many of the pathophys- iological aspects of human COPD, the effects of AZD9668 were eval- uated in guinea pigs exposed to the tobacco smoke from Kentucky R21 cigarettes (2 h, nose-only exposure) once daily, 5 days a week for a duration of 24 weeks.Groups of six to seven female Hartley strain guinea pigs (Charles River Canada), initially weighing approximately 350 g, were exposed to either air (control animals) or smoke. The smoking apparatus and details of the exposure methods have been described previously (Churg et al., 2007). AZD9668 (100 mg/kg and 1 ml/kg) or vehicle only (corn oil) was given orally by gavage according to prophylactic and therapeutic dosing regimens. In the prophylactic group, AZD9668 was given once daily 45 min before each smoke exposure for the duration of the model. In the therapeutic group, animals were treated with vehicle 45 min before smoke exposure for the first 12 weeks and then transferred to oral AZD9668 (as described above) for the final 12 weeks. At 24 weeks, the animals were euthanized, and BAL and tissue samples were removed for analysis of inflammatory indices and structural changes as described previously (Churg et al., 2007).

Statistical Analysis. Data from the acute animal models are represented as mean values ± S.E.M. × 2. Group sizes were decided by 80% power calculation. Statistical significance was calculated using a one-sided Student’s t test for a decreasing effect using pooled interanimal variability from the ANOVA.
In the chronic smoke model (data represented as mean ± S.E.M.) differences between air, smoked vehicle, and smoke plus AZD9668 groups were analyzed by ANOVA followed by Tukey’s post test for multiple comparisons made using Systat (Systat Software, Inc., San Jose, CA).

Results

Binding Kinetics. AZD9668 had a high binding affinity for human NE (KD = 9.5 nM) and potently inhibited NE activity (Table 1; Fig. 2). The calculated pIC50 (IC50) and Ki values for AZD9668 for human NE were 7.9 (12 nM) and 9.4 nM, respectively. Compared with and in contrast with ONO6818, an inhibitor that binds covalently to NE, AZD9668 exhibited a more rapid association and dissociation rate, and its interaction with NE was fully reversible.

Potency, Selectivity, and Species Crossover. AZD9668 was at least 600-fold more selective for human NE compared with other serine proteases (Table 2). Compared with reference NE inhibitors, AZD9668 generally showed greater specificity than ONO6818 or sivelestat for NE over other neutrophil-derived serine proteases, particularly Pr-3 and pancreatic elastase (Table 3).

In a panel of 319 in vitro radioligand-binding, ion-channel, and enzyme assays, significant inhibition (more than 50% inhibition at 10 µM) by AZD9668 was detected at only three targets, the adenosine transporter (IC50, 3.2 µM), noradren- aline transporter (7.2 µM), and the cholecystokinin receptor (4.6 µM) (data not shown for other targets in this panel). AZD9668 showed good crossover potency for NE from other species (Table 4).
Potency in Whole-Blood and Cell-Based Assays. The pIC50 (IC50) values for the whole-blood, cell-associated, and explosive-release assays were 7.36 (44 nM), 7.32 (48 nM), and 7.30 (50 nM), respectively (Table 5).

Activity of AZD9668 in Acute In Vivo Models. In the acute lung injury model, instillation of human NE into the trachea of mice increased BAL hemoglobin but not desmosine or hydroxyproline (data not shown). Oral administration of AZD9668 at doses >1.5 mg/kg was associated with a signif- icant reduction in human NE-induced hemoglobin levels (Fig. 3). In the rat, AZD9668 inhibited the increase in NE- induced BAL hydroxyproline in a dose-dependent manner with significant inhibition at and above 2.5 mg/kg. It also inhibited the increase in BAL desmosine at a dose of 10 mg/kg (Fig. 4).

Fig. 2. BIAcore assay of AZD9668 binding to human NE. Sensorgram shows the association and dissociation kinetics of AZD9668 and ONO6818 with immobilized human NE. The y-axis indicates the extent of interaction in BIAcore refractive units (RU), and the x-axis is time in s. After equilibration with running buffer (phosphate buffer, pH 7.4), AZD9668 or ONO6818 were injected over the immobilized enzyme at a flow rate of 50 µl/min and the association rate was deter- mined. After 1 min, running buffer was applied to the surface and the dissociation rate was determined over 5 min. The rate of complex (AB) formation and resultant on-rate (kon), off-rate (koff), and KD (koff/kon) were determined as described under Materials and Methods. Data represent one of two separate experiments.

Activity of AZD9668 in the Guinea Pig Chronic Smoke Model. In a model of chronic smoke-induced inflam- mation and emphysema in guinea pigs, AZD9668 given orally was equally effective at reducing signs of inflammation and structural change whether given prophylactically or thera- peutically. With both treatments AZD9668 prevented smoke- induced increases in lavage neutrophils and macrophages at the 6-month sacrifice time. Likewise, both treatments com- pletely prevented airspace enlargement (emphysema) and small airway remodeling (Table 6).There were no drug-related adverse effects in any of the in vivo models.

Discussion

The results of the studies presented show that AZD9668 is a specific, potent, and rapidly reversible inhibitor of human NE. In accordance with the described pathophysiological role of NE in COPD and other respiratory conditions, AZD9668 has the potential to inhibit the inflammatory burden and associated insidious decline in lung function in these condi- tions in a formulation that is convenient and well tolerated. AZD9668 differs from other small-molecule NE inhibitors, as exemplified by sivelestat, in several respects. First, the en- zyme affinity (Ki) of AZD9668 was approximately 10-fold lower than that reported for sivelestat (Kawabata et al., 1991). Second, the pharmacokinetic properties of AZD9668 make it suitable for oral dosing; sivelestat is administered by intravenous infusion. Third, its interaction with NE is rapid.

The association constant for AZD9668 (Kon 7.0 × 106 M—1 s—1) is comparable with the endogenous inhibitor, α1-AT, and therefore likely to be sufficient to inhibit elastase activity in vivo at sites of inflammation, where α1-AT may have lost its serine protease inhibitor activity because of oxidative inacti- vation (Beatty et al., 1980). Fourth, the rapid dissociation of AZD9668 from its target NE (i.e., its reversibility) contrasts with other covalent NE inhibitors, which bind to the enzyme covalently and have the potential to form drug-protein ad- ducts. These adducts may be associated with drug accumu- lation, hypersensitivity, and increased risk of toxicity (Zhou et al., 2005), which limit their long-term use in the clinic.

Fig. 3. Effects of AZD9668 on human NE-induced acute lung injury in mice. Dose response for AZD9668 measuring BAL hemoglobin is shown. Values are mean ± S.E.M. × 2 (n = 6 –15 animals/group). Statistical significance was cal- culated using a one-sided Student’s t test for a decreasing effect using pooled interanimal variability from the ANOVA. †††, P < 0.001 versus control; **, P < 0.01 versus human NE; ***, P < 0.001 versus human NE.

Fig. 4. Effects of AZD9668 on human NE-induced acute lung injury in rats. Dose versus response data for AZD9668 are shown, measuring breakdown products in BAL: hy- droxyproline (left) and desmosine (right). Values are mean ± S.E.M. × 2 (n = 4 –10 animals/group). Statistical significance was calculated using a one-sided Student’s t test for a decreasing effect using pooled interanimal vari- ability from the ANOVA. †††, P < 0.001 versus control values; *, P < 0.05 versus human NE; **, P < 0.01 versus human NE; ***, P < 0.001 versus human NE.

Fig. 5. Effects of AZD9668 on tobacco smoke-induced air- way inflammation in mice. Dose versus response data for AZD9668 measuring BAL neutrophils (left) and IL-1β (right) are shown. Values are mean ± S.E.M. × 2 (n = 16 –19 animals/group). Statistical significance was calcu- lated using a one-sided Student’s t test for a decreasing effect using pooled interanimal variability from the ANOVA. †††, P < 0.001 versus air; *, P < 0.02 versus smoke; **, P < 0.002 versus smoke; ***, P < 0.001 versus smoke.

The selectivity of AZD9668 for NE over other serine pro- teases, particularly Pr-3, was significantly greater than either ONO6818 (another NE inhibitor) or sivelestat. Whether the potential safety advantages associated with high selectivity oc- cur at the expense of efficacy is as yet unknown. Pr3 and catG, the two other granule-associated serine proteases in neutro- phils, individually have the potential to contribute to human COPD by multiple mechanisms often with significant overlap to NE (Maryanoff et al., 2010), but in contrast to NE, their role in human COPD or the relevant smoke-driven animal models has yet to be demonstrated.

The potent inhibitory activity of AZD9668 on NE in bio- chemical assays was confirmed in whole-blood and cell-based systems. In the whole-blood assay, high levels of NE are released on stimulation, which overcome the serine protease inhibitor capacity of blood. In this assay, AZD9668 inhibited zymosan-stimulated NE activity in plasma with a pIC50 of 7.36 (46 nM). This translated to an actual potency of 29 nM when adjusted for plasma protein binding. The whole-blood assay has been useful in predicting appropriate therapeutic dosing and exposure in humans. In a recent clinical trial it was shown that AZD9668 inhibited ex vivo zymosan-stim- ulated NE activity after oral dosing in human subjects (Gunawardena et al., 2010).

The cell-associated NE assay was developed to confirm inhibition of NE activity expressed specifically on the cell surface rather than the soluble form of the enzyme. Studies have suggested that the surface-bound catalytically active enzyme may be as important as the soluble form in mediating pathological effect (Young et al., 2007; Owen, 2008) and may be more resistant to endogenous antiprotease inhibitors (Owen et al., 1995). However, direct evidence in humans for a specific pathophysiological role of the cell-associated en- zyme is still lacking (Tang et al., 2004). Nevertheless, similar to its activity in whole blood, AZD9668 proved to be a potent inhibitor of cell-associated NE with a pIC50 of 7.31 (48 nM). Aside from the NE activity expressed on the surface of activated cells, quantal release of soluble NE has also been reported (Liou and Campbell, 1995). The rapid, localized release of high concentrations of NE that result may lead to a zone of destruction around the neutrophil (Aboussouan and Stoller, 2008) where interaction with physiological sub- strates and inhibitor may occur under nonequilibrium condi- tions. In this context, most efficient inhibition will probably be achieved with inhibitors with rapid enzyme association rates (e.g., α1-AT). In the explosive assay, AZD9668, with its rapid association constant, inhibited NE activity potently, with an pIC50 of 7.30 (50 nM).

As well as its effects in vitro, AZD9668 reduced levels of human NE-induced BAL hemoglobin in a mouse model of acute lung injury and reduced human NE-induced BAL des- mosine and hydroxyproline in an equivalent rat model. We have not observed any significant increase in BAL hydroxy- proline and desmosine after instillation of human NE in the mouse, perhaps because of sample volumes being insufficient for analysis.The better efficacy observed in the mouse reflects a lower AZD9668 plasma protein binding compared with rat and therefore improved exposure of the drug after a given oral dose (data not shown).

Desmosine is an atypical amino acid that cross-links and provides structure to elastin; in COPD, urinary desmosine is regarded as a biomarker of lung injury and levels are increased according to disease severity and the rate of decline of lung function (Gottlieb et al., 1996; Luisetti et al., 2008). AZD9668 has also been shown to significantly reduce urinary desmosine levels after 4 weeks of oral administration to patients with CF (Elborn et al., 2011). By measuring effects on markers relevant to human lung disease, these data show that AZD9668 inhibits NE-mediated lung injury. Likewise, 6 weeks of treatment with 2-(2-thiophencarboxythio)-N-[dihydro-2(3H)-thiophenone-3-yl]- propionamide (MR889), a synthetic cyclic thiolic NE inhibitor, reduced desmosine levels in a subgroup of patients with COPD with short disease duration (Luisetti et al., 1996).

The role of inflammatory cytokines (e.g., IL-6 and IL-8) in COPD is well established, particularly during disease exac- erbations (Bhowmik et al., 2000). In the in vivo model of acute smoke-induced lung inflammation, treatment with AZD9668 was associated with a significant reduction of neu- trophils and IL-1β in BAL. These results are consistent with previous studies demonstrating that NE has the capacity to stimulate diverse proinflammatory signaling pathways through multiple mechanisms (Be´dard et al., 1993; Wiedow and Meyer-Hoffert, 2005).
The effects of AZD9668 in the chronic smoke model were consistent with those observed in the more acute models. Both prophylactic and therapeutic treatment prevented increases in lavage inflammatory cells and structural changes in the form of emphysema and small airway remodeling. We had previously reported that another NE inhibitor, L-prolinamide [N-(4- methoxybenzoyl)-L-valyl]-N-[3,3,3-trifluoro-(1S)-(1-methyl- ethyl)-2-oxopropyl]-L-prolinamide (ZD0892) (Wright et al., 2002) reduced lavage inflammatory cells and partially pre- vented emphysema in a similar model (airway remodeling was not examined) when given prophylactically, but was not effec- tive in preventing emphysema when started at 4 months of smoke exposure, although it did reduce inflammatory cells with this regime. Whether these differences are an effect of dose, timing, or specificity and fast onset mechanism of AZD9668 compared with ZD0892 is not clear. It is interesting that AZD9668 was able to completely prevent emphysema even though it does not inhibit Pr-3 or catG (at least in humans), suggesting that those proteases may not play a role in the genesis of emphysema. This is also the first demonstration that a NE inhibitor can prevent small airway remodeling, an obser- vation of clinical importance, because there is increasing evi- dence that small airway remodeling is a major cause of airflow obstruction in COPD.

Taken together, these findings suggest that NE inhibition has the potential to have a beneficial effect in inflammatory lung diseases by modulating inflammatory mediators and lung destruction. In a recent exploratory study in patients with bronchiectasis, 4 weeks of treatment with AZD9668 significantly improved some lung function measures and re- duced the inflammatory mediators IL-6 and regulated upon activation normal T cell expressed and secreted (Stockley et al., 2010).

To conclude, AZD9668 is a potent, selective, and reversible inhibitor of human NE with good efficacy in preclinical mod- els. Its pharmacological profile suggests that it represents a significant advance versus previous NE inhibitors, and it has the potential to be effective for neutrophil-driven inflamma- tory lung diseases, such as bronchiectasis and COPD.

Acknowledgments

We thank Dr. Henrik Lindberg and Dr. Claes Lindberg for assis- tance with the mass spectroscopic analysis of desmosine and hy- droxyproline and Dr. Nicky French from Complete Medical Commu- nications (funded by AstraZeneca) for editorial support.

Authorship Contributions

Participated in research design: Stevens, Ekholm, Gra¨ nse, Jungar, Ottosson, Churg, Wright, Lal, and Sanfridson.
Conducted experiments: Stevens, Ekholm, Gra¨ nse, Lindahl, Kozma, Jungar, Ottosson, Falk-Håkansson, Churg, and Wright.

Performed data analysis: Stevens, Ekholm, Gra¨ nse, Kozma, Jun- gar, Ottosson, Churg, Wright, Lal, and Sanfridson.
Wrote or contributed to the writing of the manuscript: Stevens, Ekholm, Gra¨ nse, Lindahl, Kozma, Jungar, Ottosson, Falk-Håkans- son, Churg, Wright, Lal, and Sanfridson.

References

Aboussouan LS and Stoller JK (2008) a1-Antitrypsin Deficiency, in Chronic Obstruc- tive Pulmonary Disease (Stockley RA, Rennard SI, Rabe K, and Celli B eds) pp 447– 459, Blackwell Publishing Limited, Boston, MA.
Beatty K, Bieth J, and Travis J (1980) Kinetics of association of serine proteinases with native and oxidized α-1-proteinase inhibitor and α-1-antichymotrypsin. J Biol Chem 255:3931–3934.
Be´dard M, McClure CD, Schiller NL, Francoeur C, Cantin A, and Denis M (1993) Release of interleukin-8, interleukin-6, and colony-stimulating factors by upper airway epithelial cells: implications for cystic fibrosis. Am J Respir Cell Mol Biol 9:455– 462.
Bergin DA, Greene CM, Sterchi EE, Kenna C, Geraghty P, Belaaouaj A, Belaaouaj A, Taggart CC, O’Neill SJ, and McElvaney NG (2008) Activation of the epidermal growth factor receptor (EGFR) by a novel metalloprotease pathway. J Biol Chem 283:31736 –31744.
Betsuyaku T, Nishimura M, Yoshioka A, Takeyabu K, Miyamoto K, and Kawakami Y (1996) [Neutrophil elastase and elastin-derived peptides in BAL fluid and emphysematous changes on CT scans]. Nihon Kyobu Shikkan Gakkai Zasshi 34(Suppl):69 –74.
Bhowmik A, Seemungal TA, Sapsford RJ, and Wedzicha JA (2000) Relation of sputum inflammatory markers to symptoms and lung function changes in COPD exacerbations.Thorax 55:114 –120.
Caldwell RA, Boucher RC, and Stutts MJ (2005) Neutrophil elastase activates near-silent epithelial Na+ channels and increases airway epithelial Na+ trans- port. Am J Physiol Lung Cell Mol Physiol 288:L813–L819.
Chan SC, Shum DK, and Ip MS (2003) Sputum sol neutrophil elastase activity in bronchiectasis: differential modulation by syndecan-1. Am J Respir Crit Care Med 168:192–198.
Cheng Y and Prusoff WH (1973) Relationship between the inhibition constant (K1) and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction. Biochem Pharmacol 22:3099 –3108.
Churg A, Wang R, Wang X, Onnervik PO, Thim K, and Wright JL (2007) Effect of an MMP-9/MMP-12 inhibitor on smoke-induced emphysema and airway remodelling in guinea pigs. Thorax 62:706 –713.
Damiano VV, Tsang A, Kucich U, Abrams WR, Rosenbloom J, Kimbel P, Fallahnejad M, and Weinbaum G (1986) Immunolocalization of elastase in human emphyse- matous lungs. J Clin Invest 78:482– 493.
Elborn JS, Perrett J, Forsman-Semb K, Marks-Konczalik J, Gunawardena K, and Entwistle N (2011) Effect of the oral neutrophil elastase inhibitor, AZD9668, in patients with cystic fibrosis, at the 107th Annual Conference of the American Thoracic Society; 2011 May 13–18; Denver, CO. May 13–18, 2011. American Thoracic Society, New York.
Gottlieb DJ, Stone PJ, Sparrow D, Gale ME, Weiss ST, Snider GL, and O’Connor GT (1996) Urinary desmosine excretion in smokers with and without rapid decline of lung function: the Normative Aging Study. Am J Respir Crit Care Med 154:1290 – 1295.
Gunawardena K, Gullstrand H, and Perrett J (2010) Safety, tolerability and phar- macokinetics of AZD9668, an oral neutrophil elastase inhibitor, in healthy subjects and patients with COPD, at the European Respiratory Society Annual Congress; 2010 Sept 18 –22; Barcelona, Spain. Abstract 1163, 203s. European Respiratory Society, Lausanne, Switzerland.
Hayakawa M, Katabami K, Wada T, Sugano M, Hoshino H, Sawamura A, and Gando S (2010) Sivelestat (selective neutrophil elastase inhibitor) improves the mortality rate of sepsis associated with both acute respiratory distress syndrome and dis- seminated intravascular coagulation patients. Shock 33:14 –18.
Hogg JC, Chu F, Utokaparch S, Woods R, Elliott WM, Buzatu L, Cherniack RM, Rogers RM, Sciurba FC, Coxson HO, et al. (2004) The nature of small-airway obstruction in chronic obstructive pulmonary disease. N Engl J Med 350:2645– 2653.
Kawabata K, Suzuki M, Sugitani M, Imaki K, Toda M, and Miyamoto T (1991) ONO-5046, a novel inhibitor of human neutrophil elastase. Biochem Biophys Res Commun 177:814 – 820.
Luisetti M, Ma S, Iadarola P, Stone PJ, Viglio S, Casado B, Lin YY, Snider GL, and Turino GM (2008) Desmosine as a biomarker of elastin degradation in COPD: current status and future directions. Eur Respir J 32:1146 –1157.
Luisetti M, Sturani C, Sella D, Madonini E, Galavotti V, Bruno G, Peona V, Kucich U, Dagnino G, Rosenbloom J, et al. (1996) MR889, a neutrophil elastase inhibitor, in patients with chronic obstructive pulmonary disease: a double-blind, random- ized, placebo-controlled clinical trial. Eur Respir J 9:1482–1486.
Maryanoff BE, de Garavilla L, Greco MN, Haertlein BJ, Wells GI, Andrade-Gordon P, and Abraham WM (2010) Dual inhibition of cathepsin G and chymase is effective in animal models of pulmonary inflammation. Am J Respir Crit Care Med 181:247–253.
Morrison HM, Kramps JA, Burnett D, and Stockley RA (1987) Lung lavage fluid from patients with α1-proteinase inhibitor deficiency or chronic obstructive bron- chitis: anti-elastase function and cell profile. Clin Sci (Lond) 72:373–381.
O’Donnell RA, Peebles C, Ward JA, Daraker A, Angco G, Broberg P, Pierrou S, Lund J, Holgate ST, Davies DE, et al. (2004) Relationship between peripheral airway dysfunction, airway obstruction, and neutrophilic inflammation in COPD. Thorax 59:837– 842.
Owen CA (2008) Leukocyte cell surface proteinases: regulation of expression, func- tions, and mechanisms of surface localization. Int J Biochem Cell Biol 40:1246 – 1272.
Owen CA, Campbell MA, Sannes PL, Boukedes SS, and Campbell EJ (1995) Cell surface-bound elastase and cathepsin G on human neutrophils: a novel, non- oxidative mechanism by which neutrophils focus and preserve catalytic activity of serine proteinases. J Cell Biol 131:775–789.
Rubin H (1996) Serine protease inhibitors (SERPINS): where mechanism meets medicine. Nat Med 2:632– 633.
Sinha S, Watorek W, Karr S, Giles J, Bode W, and Travis J (1987) Primary structure of human neutrophil elastase. Proc Natl Acad Sci U S A 84:2228 –2232.
Stockley RA, Snell N, Perrett J, and Gunawardena K (2010) Efficacy and safety of AZD9668, an oral neutrophil elastase inhibitor, in idiopathic bronchiectasis, at the European Respiratory Society Annual Congress; 2010 Sept 18 –22; Barcelona, Spain. Abstract 375, 38s. European Respiratory Society, Lausanne, Switzerland. Stolk J, Veldhuisen B, Annovazzi L, Zanone C, Versteeg EM, van Kuppevelt TH, Berden JH, Nieuwenhuizen W, Iadarola P, and Luisetti M (2005) Short-term variability of biomarkers of proteinase activity in patients with emphysema asso-
ciated with type Z α-1-antitrypsin deficiency. Respir Res 6:47.
Tang M, Gu YJ, Wang WJ, Xu YP, and Chen CZ (2004) Effect of cardiopulmonary bypass on leukocyte activation: changes in membrane-bound elastase on neutro- phils. Perfusion 19:93–99.
Tremblay GM, Vachon E, Larouche C, and Bourbonnais Y (2002) Inhibition of human neutrophil elastase-induced acute lung injury in hamsters by recombinant human pre-elafin (trappin-2). Chest 121:582–588.
van der Vaart H, Postma DS, Timens W, and ten Hacken NH (2004) Acute effects of cigarette smoke on inflammation and oxidative stress: a review. Thorax 59:713– 721.
Voynow JA, Young LR, Wang Y, Horger T, Rose MC, and Fischer BM (1999) Neutrophil elastase increases MUC5AC mRNA and protein expression in respira- tory epithelial cells. Am J Physiol Lung Cell Mol Physiol 276:L835–L843.
Weldon S, McNally P, McElvaney NG, Elborn JS, McAuley DF, Wartelle J, Belaa- ouaj A, Levine RL, and Taggart CC (2009) Decreased levels of secretory leucopro- tease inhibitor in the Pseudomonas-infected cystic fibrosis lung are due to neutro- phil elastase degradation. J Immunol 183:8148 – 8156.
Wiedow O and Meyer-Hoffert U (2005) Neutrophil serine proteases: potential key regulators of cell signalling during inflammation. J Intern Med 257:319 –328.
Wright JL, Farmer SG, and Churg A (2002) Synthetic serine elastase inhibitor reduces cigarette smoke-induced emphysema in guinea pigs. Am J Respir Crit Care Med 166:954 –960.
Young RE, Voisin MB, Wang S, Dangerfield J, and Nourshargh S (2007) Role of neutrophil elastase in LTB4-induced neutrophil transmigration in vivo assessed with a specific inhibitor and neutrophil elastase-deficient mice. Br J Pharmacol 151:628 – 637.
Yurewicz EC and Zimmerman M (1977) Cytochalasin B-dependent release of azuro- phil granule enzymes from human polymorphonuclear leukocytes. Inflammation 2:259 –264.
Zhou S, Chan E, Duan W, Huang M, and Chen YZ (2005) Drug bioactivation, covalent binding to target proteins and toxicity relevance. Drug Metab Rev 37:41– 213.