Effects of selective phosphodiesterase inhibitors on the polymorphonuclear leukocyte respiratory burst Christopher P. Nielson, MD,* Robert E. Vestal, MD,* Robert J. Sturm, PhD,** and Richard Heaslip, PhD** Boise, Idaho, Seattle, Wash., and Princeton, N.J.

Modulation of the human polymorphonuclear leukocyte (PMN) respiratory burst by selective cyclic 3',5' adenosine monophosphate (cAMP)phosphodiesterase (PDE) inhibitors was studied with respect to PDE isozyme characteristics. Zaprinast, an inhibitor of a cyclic guanosine monophosphate (cGMP)-specific PDE (PDE I), at concentrations up to 100 tzmol/ L, had no significant effect on the respiratory burst. Milrinone and imazodan, inhibitors of cAMP-metabolizing, cGMP-sensitive PDE (PDE III), reduced the respiratory burst to 60% of control magnitude but only had significant effects when they were introduced at high (100 Izmol / L ) concentrations. In contrast, rolipram and RO 20-1724, inhibitors of a cAMP-metabolizing, cGMP-insensitive PDE (PDE IV), had significant effects at low concentrations (0.1 txmol/L) and caused marked reduction of the respiratory burst at higher concentrations (25% of control at 10 tzmol/L). The selective PDE 1V inhibitors significantly potentiated PMN inhibition by isoproterenol. Diethylaminoethyl (DEAE)-Sepharose chromatography demonstrated a predominant PDE isozyme with high affinity and selectivity for cAMP that was insensitive to cGMP and was completely inhibited by rolipram, a PDE IV inhibitor. These results are consistent with the conclusion that the PMN respiratory burst is inhibited by an elevation of cAMP induced by PDE IV inhibition. ( J ALLERGY CLIN 1MMUNOL 1990;86:801-8.)

The PMN responds to extracellular stimuli with a respiratory burst that generates toxic oxygen metabolites.~ Although the PMN is a critical cell in host defense, excessive or inappropriate PMN response may cause painful inflammation, bronchial hyperreactivity, and tissue injury. Consequently, pharmacologic suppression of the PMN respiratory burst could be important in the treatment of such diverse pathologic processes as arthritis, asthma, and myocardial infarction. 2' 3

From the *Clinical Pharmacologyand GerontologyResearch Unit, Veterans AdministrationMedical Center, Boise, Idaho, *Division of Gerontologyand Geriatric Medicine, Departmentof Medicine, University of Washington, Seattle, Wash., and **Wyeth-Ayerst Research, Division of Im_munopharmacology,Princeton, N.J. Supported in part by the Department of Veterans Affairs and National Institutes of Health Grant DK 40325. Received for publication Feb. 13, 1990. Revised June 26, 1990. Accepted for publication June 29, 1990. Reprint requests: Christopher Nielson, MD, Research Service (151), VeteransAffairs Medical Center, 500 WestFort St., Boise, ID 83702. 1/1/23572

Abbreviations used cAMP: Cyclic 3',5' adenosine monophosphate FMLP: N-formyl-methionyl-leucyl-phenylalanine PDE: Phosphodiesterase PMA: Phorbol myristate acetate PMN: Polymorphonuclear leukocyte IC~o: Drug concentration causing 50% maximal inhibition PBS: Phosphate-buffered saline DEAE: Diethylaminoethyl cGMP: Cyclic guanosine monophosphate K~: Michaelis-Menten constant

Agents that increase PMN cAMP, including 13adrenoceptor agonists and methylxanthines, inhibit the respiratory burst 4-7 and reduce inflammatory mediator release. Both [3-adrenoceptor agonists and methylxanthines are of major importance in the treatment of asthma, a therapeutic effect that may be at least partially mediated by inhibition o f PMN actb vation. 7 However, because these agents have actions

801

802 Nielson et al.

J. ALLERGYCLIN.IMMUNOL NOVEMBER 1990

2500 2400 2300

~2oo

F"I Cont r OI

1 0o/

2100

m l O 0 nM ISO O 5uM Roliprarn

2000

1900

OISO

+ Rolipram

1800 1700

1600

1400 1300

1200

1100 1000 900 800

~'~.~

700 600 500

_ 0

_ 5

10

15

20

25

TIME (MINUTES)

FIG. 1. Representative experiment (one of five) illustrating respiratory burst after PMN activation with 0,2 t~mot/L of calcium ionophore A23187 in the presence of 100 nmol/L of isoproterenol, 5 t~mol/L of rolipram, or both drugs.

on multiple cell types, their use is often limited by adverse cardiovascular and neurologic effects. CAMP PDE occurs as multiple isozymes, and potent agents that selectively inhibit the PDE I (cGMP specific), PDE III (cAMP specific, cGMP inhibited), and PDE IV (cAMP specific, cGMP insensitive) isozymes have been identified. 8, 9 Since the distribution and functional importance of specific isozymes is variable among tissues, 1~ selective PDE inhibitors may have fewer adverse effects than nonselective inhibitors, such as theophylline. The purpose of this study was to characterize the effects of selective PDE inhibitors on the PMN respiratory burst, identify the isozymes of PDE present in PMNs, and determine whether selective PDE inhibitors in combination with [3-adrenoceptor agonists may reduce superoxide anion generation. Pharmacologic modulation of the PMN respiratory burst could be of clinical value in inflammatory disease.

MATERIAL AND METHODS PMN separation For studies of PMN function, blood samples were obtained from healthy volunteers, aged 20 to 45 years, after informed consent was obtained. The study was approved by the Human Subjects Committee of the University of Washington. No subject was using any medication. Cells were isolated from venous blood anticoagulated with 10

U/ml of heparin. To sediment the red blood cells, 30 ml of blood was added to 30 ml of dextran containing 1 mg/ml ethylenediaminetetraaceticacid. After 20 minutes, the leukocyte-rich plasma was layered onto 3 ml FicollHypaque (SpG 1.077)/1 PMNs were isolated with centrifugation (200 g for 20 minutes) and hypotonic (10 seconds in distilled water) lysis of erythrocytes. PMNs were stored in plasma from the same donor at 4 ~ C before use. Immediately before each experiment, PMNs were removed from plasma with centrifugation at 200 g for 10 minutes, washed, and resuspended in phosphate buffer. The final preparation was >95% PMNs, and PMNs were 95% viable by trypan blue exclusion. For studies of isolated PMN PDE, leukocyte-enriched serum samples were obtained from healthy male subjects with a Haemonetics (Bentley Laboratories, Inc., Irvine, Calif.) model 30 + blood processor. Samples were centrifuged at 35 g for 10 minutes at 20~ C, and the top plateletrich layer was discarded (approximately 10%). The remaining specimen was centrifuged at 400 g for 10 minutes, the pellet was resuspended in Ca +§ and Mg § free Hanks' balanced salt solution, and PMNs were isolated with FicollHypaque sedimentation" and hypotonic lysis of erythrocytes. PMNs were washed one time, and the final pellet (>95% PMNs) was frozen at - 2 0 ~ C before use for PDE studies.

Evaluation of PMN respiration burst After cell suspension in phosphate buffer, PMNs were activated by addition of chemotactic peptide FMLP, calcium ionophore A23187, or the protein kinase C agonist, PMA. Lucigenin (10,10'-dimethyl-bis-9,9'-biacridinium nitrate) was used to detect superoxide anion generation. The luminescence response from 2 • 105 cells in 2 ml of buffer at 37~ C was measured at 2-minute intervals with a Picolight luminometer (Packard Instrument Co.), model 6500. Lucigenin-dependent luminescence was completely inhibited by superoxide dismutase (100 U/ml) and therefore appeared to correlate with superoxide anion generation, as previously reported. 12Neither rolipram nor RO 20-1724 inhibited lucigenin-dependent luminescence induced by xanthine (500 ~mol/L) and xanthine oxidase (0.025 units). The PMN respiratory burst was evaluated during an interval of 4 to 30 minutes after PMN activation (area under curve). The magnitude of respiratory burst from each sample was standardized as a percentage of the response from a control sample. Comparisons of drug effects were paired with PMNs from the same blood sample with activation and measurements performed in parallel under identical experimental conditions. Experiments were usually repeated five times with specimens from different donors. Dose-response curves were evaluated with computer-assisted nonlinear curve fitting to the logistic equation TM i, for estimation of slope, ICs0, and maximal response.

Phosphodiesterase isolation and evaluation Isozymes of PDE were purified and assayed according to a modification of the method of Thompson et al. 15 PMNs (6 x 109) were hand homogenized in 35 ml of ho-

VOLUME86

Phosphodiesterase inhibition and leukocytes

803

NUMBER 5

mogenization buffer (10 mmol/L of Tris HCL, 5 mmol/L of MgCL2, 4 mmol/L of ethylene glycol-bis-(betaaminoethyl ether)-N,N', tetraacetic acid (EGTA), 5 mmol/L of [3-mercaptoethanol, 1% Triton X-100, 1 Ixmol/L of leupeptin, 1 ixmol/L of pepstatin, and 100 txmol/L of phenylmethyl sulfonyl fluoride, pH 7.8) and centrifuged at 25,000 g at 4 ~ C for 30 minutes. The supernatant was applied to a 40 / 2 cm DEAE-Sepharose CL-6B column that had been washed in Triton-free homogenization buffer. Isozymes of PDE were eluted from the column (80 ml/hr) with a linear 0 to 1.0 mol/L gradient of sodium acetate in elution buffer (10 mmol/L of Tris HCL, 5 mmol/L of MgCL2, 2 mmol/L of ethylene glycol-bis(beta-aminoethyl ether)-N,N', tetraacetic acid, 5 mmol/L of 13-mercaptoethanol, 1% Triton X-100, 0.1 Ixmol/L of leupeptin, 0.1 ~mol/L of pepstatin, and 50 txmol/L of phenylmethyl sulfonyl fluoride, pH 7.8), and 7.5 ml fractions were collected. Each fraction was assayed in duplicate for cAMP- and cGMP-metabolizing PDEs with 1 ixmol/L of 3H-cyclic nucleotide substrate by the Dowex l-X8 affinity-column method.I~ Fractions containing similar PDE activities were pooled and concentrated to 15% volume with an Amicon (Amicon Corp., Danvers, Mass.) filtration system. Ethylene glycol (30% vol/vol) was added, and samples were stored at - 2 0 ~ C. Inhibition of PDE by isozyme-selective inhibitors was assessed in 10 mmol/L of Tris HC1, 5 mmol/L of MgCI2, and 1 mmol/L of [3-mercaptoethanol by measuring inhibition of cyclic nucleotide metabolism at a 1 txmol/L of substrate concentration under conditions in which < 10% of total substrate was metabolized. All ICs0s (50% inhibitory concentrations) were calculated by regression analysis of percent-inhibition data. Reagents and materials. Dulbecco's PBS was prepared with 1 mg/ml of glucose, 1 mmol/L of MgC12, and 1 mmol/L of CaCL2. Calcium ionophore A23187, FMLP, and PMA were dissolved in dimethylsulfoxide and diluted 1 : 100 in PBS before cell stimulation. Lucigenin was dissolved in buffer. Isoproterenol was initially prepared in distilled water with 1 mg/ml of metabisulfite to prevent oxidation. Isoproterenol was diluted 1 : 100 in PBS immediately before each experiment. Diluents were included in all samples of any study at equal concentrations. No effects on PMN function were observed at the concentrations of diluents used. RO 20-1724 was purchased from Biomol Research Laboratories, Plymouth Meeting, Pa. Enprofylline (AB Draco/Astra, Lund, Sweden), zaprinast (M&B 22948; May and Baker, Ltd., Dagenham, England), milrinone (WIN 47,203-2; Sterling-Winthrop, Rensselaer, N.Y.), imazodan (CI-914; Parke-Davis, Ann Arbor, Mich.), and rolipram (ZK 62,711; Schering AG, Berlin, West Germany) were gifts from the respective manufacturers. Leukocyte-enriched serum used for PDE isolation was purchased from Biological Specialties Corp., Lansdale, Pa. All other materials were purchased from Sigma Chemical Co., St. Louis, Mo. RESULTS In control samples without PDE inhibitors, superoxide anion generation after PMN activation by 0.2

CONCENTRATION [] 1 uM 9 [] i

THEO. ENPRO. ZAPRIN. MILRIN. iMAZ. PDE INHIBITOR

10 u M

[]

100

uM

ROL|P+RO20-1724

FIG. 2. Effects of PDE inhibitors on respiratory burst induced by calcium ionophore A23187. PDE inhibitors at concentrations indicated were introduced at the time of respiratory burst activation by calcium ionophore A23187. Lucigenin-dependent luminescence was integrated during 26 minutes and is represented as a percentage of the simultaneously measured control sample. Data indicate means-4-SEM of specimens from four subjects (*p < 0.05).

txmol/L compound ionophore A23187 rapidly increased to a peak at 2 to 4 minutes and gradually resolved during 30 to 60 minutes. A representative study demonstrating the respiratory burst and inhibitory effects of isoproterenol and rolipram is illustrated in Fig. 1. The rapid initial increase in oxygen metabolite generation was effectively suppressed by isoproterenol, whereas later phases of the respiratory burst were markedly inhibited by rolipram. PDE inhibitors introduced at the beginning of the respiratory burst reduced oxygen metabolite generation, as illustrated in Fig. 2. The respiratory burst in PMN exposed to each agent was calculated as a percentage of the respiratory burst in an aliquot of PMN activated and analyzed in parallel but not exposed to PDE inhibitor. Only the selective PDE IV inhibitors, rolipram and RO 20-1724, caused significant inhibition at concentrations < 1 0 0 ixmol/L. The nonspecific PDE inhibitors, theophylline and enprofylline (100 txmol/L), inhibited to 65% --+ 7% control and 50% - 6% control, respectively. Zaprinast (PDE I) (100 txmol/L) caused an increase in the respiratory burst to 126% _ 9% control (not significant, p > 0.05). High concentrations (100 ixmol/L) of mil-

804

N i e l s o n e t al.

J. ALLERGY CLIN. IMMUNOL. NOVEMBER 1990

110

14C

105

I

100.

I

9s

1

90.

[~Ro2O-

80. -I

0

1

m Rolipram

85,

1724 E I-Z 0 Q

rs. 70.

o C,~ 6 5 , I-Z 60. I.g O ~ 55. tU a. 50.

Z LU a: LU O.

45. 40. 35, 30

EARLY LATE ( 2 - 4 MIN) ( 2 0 - 2 6 MIN) NO INCUBATION

25 201 -9

0 -8 LOG

I I -7 -B CONCENTRATION

I -5 (M)

I -4

FIG. 3. Effects of rolipram and Ro 20-1724 on respiratory burst induced by chemotactic peptide (1 ixmol/L of FMLP). PDE inhibitors were introduced at the time of PMN stimulation by FMLP, and magnitude of the respiratory burst was integrated during 26 minutes. Curves were generated with computer-assisted nonlinear fitting to the logistic equation. Data represent the means -+ SEM of samples from five subjects.

rinone and imazodan (PDE III) inhibited to 81% ___ 5% and 84% __+ 5% of control, respectively (mean _ SEM; n = 4; p < 0.05). In contrast, concentrations of rolipram and RO 20-1724 >1 p.mol/L caused significant inhibition (n = 4; p < 0.05). Stimulation of the respiratory burst by either calcium ionophore or the chemotactic peptide, FMLP, has previously been demonstrated to be sensitive to regulation by agents that increase cAMP. 4' 6. 16 Consistent with these studies, the respiratory burst, when it was stimulated by FMLP, was inhibited by PDE inhibitors in a manner similar to PMN stimulated by calcium ionophore A23187. Both rolipram and Ro 201724 caused potent inhibition with ICsos of 0.2 0.07 txmol/L and 1.1 - 0.6 p~mol/L, and maximal inhibition to 30% ___ 4% control and 28% ___ 10% control, respectively (Fig. 3, computer-assisted nonlinear curve fitting; n = 5). Rolipram and RO 20-1724 had less effect than isoproterenol (Fig. 1) or the nonspecific PDE inhibitors,

I~lTheophylline [-'] Enpr o f ylline

mZaprinast

EARLY LATE ( 2 - 4 MIN) ( 2 0 - 2 6 MIN) 3 0 MIN P R E I N C U B A T I O N k~Milrinone

mRoliprarn

[]lmazodan

mlRO20-1724

FIG. 4. Effects of PDE inhibitors on respiratory burst induced by 0.2 izmol/L of calcium ionophore A23187. Incubated PMNs were exposed to either 0.5 txmol/L (rolipram or Ro 20-1724) or 50 txmol/L (other agents) PDE inhibitor during incubation and 0.005 ixmol/L (rolipram or Ro 20-1724) or 0.5 ixmol/L (other agents) during respiratory burst. Nonincubated PMNs were exposed to 1 ~mol/L (rolipram or Ro 20-1724) or 100 ixmol/L (other agents) PDE inhibitor at the time of stimulation by calcium ionophore A23187. Data represent the means _+ SEM with samples from five subjects (*p < 0.05).

theophylline and enprofylline, when these were evaluated during the initial 5 minutes after PMN activation (Fig. 4). In contrast, the selective PDE inhibitors caused profound inhibition during later periods of the respiratory burst (Fig. 4, "No incubation" bars). To allow intracellular diffusion, PMNs were preincubated with each PDE inhibitor in autologous plasma for 30 minutes at 37 ~C. The cells, plasma, and PDE inhibitor were then diluted 1:100 with PBS, and the PMNs were activated with calcium ionophore A23187. Thus, the extracellular concentration of PDE inhibitor during the respiratory burst was 1 / 100 of the concentration during preincubation. Rolipram and RO 20-1724 caused marked inhibition after PMN preincubation with low drug concentrations (0.5 p.mol/L preincubation; 0.005 Ixmol/L during respiratory burst). Other PDE inhibitors at much higher concentrations (50 p.mol/L concentration during preincubation; 0.5

VOLUME 86 NUMBER 5

P h o s p h o d i e s t e r a s e inhibition and leukocytes

805

60 0,

5O O

In