Cyclic guanosine monophosphate (GMP) plays an integral role in mediating relaxation of vascular smooth muscle. Nitrovasodilators such as endogenously released nitric oxide (NO) and the drug nitroprusside (NP) activate a soluble isoenzyme form of guanylyl cyclase to produce cGMP from GTP (1-3). cGMP induces relaxation by activating cGMP-dependent protein kinases and the phosphorylation state of several intracellular proteins including myosin light chain (4). cGMP-dependent protein kinase may activate a calcium-dependent adenosine triphosphatase (ATPase) that decreases cytosolic calcium concentration, and consequently myosin light chain kinase activity, so that vasodilation results.
Vasodilation can be induced, not only by increasing production of cGMP, but also by blocking metabolism of cGMP by cGMP-dependent (type V) phosphodiesterases (PDEs) (5-8). Zaprinast (ZAP, 2-o-propoxyphenyl-8-azapurin-6-one) is one such inhibitor (9,10). One aim was to determine whether vasorelaxation induced by submaximal concentrations of NP and ZAP is additive in intact hearts and isolated aortae. A second aim was to determine whether submaximal vasodilating concentrations of NP and ZAP, alone or together, have an effect on cardiac electrophysiology, mechanical function, and metabolism. A third was to assess whether the median inhibitory concentration (IC50) of NP on vasorelaxation is reduced by ZAP. The concentrations of NP and ZAP that individually increased coronary flow (CF) and coronary sinus O2 tension in hearts, and decreased tone in preconstricted aortic rings by 50% (IC50) were determined first. Second, the approximate IC50 concentration for each drug (when given together) was measured to assess additive or interactive effects on CF, cardiac function, and O2 utilization in intact hearts, and in aortic rings, on relaxation and cGMP production.
Approval from the institutional Animal Studies Committee was obtained before initiating this study. The investigation conforms with the Guide for the Care and Use of Laboratory Animals, published by the U.S. National Institutes of Health (NIH publication 85-23, 1995). Albino English short-haired guinea pigs were injected intraperitoneally with 10 mg of ketamine and 1,000 units of heparin and decapitated 15 min later when unresponsive to noxious stimulation.
Isolated heart studies
Our methods have been described in detail previously (11,12). After thoracotomy, the inferior and superior venae cavae were cut, and the aorta was cannulated distal to the aortic valve. Each heart was immediately perfused in a retrograde manner through the aorta with cold oxygenated modified Krebs-Ringer's solution (equilibrated with 97% O2 and 3% CO2) and then rapidly excised. All hearts were perfused at an aortic root perfusion pressure of 55 mm Hg. The perfusate, a modified Krebs-Ringer's solution, was filtered (5 μm pore size) in-line and had the following control composition in 10−3M: Na+, 137; K+, 5; Mg2+, 1.2; Ca2+, 2.5; Cl−, 134; HCO3−, 15.5; H2PO4−, 1.2; glucose, 11.5; pyruvate, 2; mannitol, 16; EDTA (ethylenediaminetetraacetic acid), 0.05; and insulin, 5 units/L. Perfusate and bath temperatures were maintained at 37.5 ± 0.1°C using a thermostatically controlled water circulator.
Left ventricular pressure (LVP) was measured isovolumetrically with a transducer connected to a thin, saline-filled latex balloon, inserted into the left ventricle through the mitral valve from a cut in the left atrium. Balloon volume was adjusted to maintain a diastolic LVP of 0 mm Hg during the initial control period, so that any increase in diastolic LVP reflected an increase in LV wall stiffness or diastolic contracture. Two pairs of bipolar electrodes were placed in each heart to monitor intracardiac electrograms, from which spontaneous atrial heart rate and atrioventricular (AV) conduction time were determined from the right atrial and ventricular beats. Coronary sinus effluent was collected by placing a cannula into the right ventricle through the pulmonary artery after ligating the venae cavae. Coronary flow (aortic inflow, CF) was measured at constant temperature and constant perfusion pressure (55 mm Hg) by a self-calibrating, in-line, ultrasonic flowmeter (Transonic T106X; Ithaca, NY, U.S.A.) placed directly into the aortic inflow line. To determine maximal CF in arrested hearts, adenosine (0.2 ml of 2 × 10−4M stock solution) was injected into the aortic root cannula during the initial control period and after the last control reading.
Coronary arterial O2 tension was measured with an intermittently self-calibrating analyzer system (Radiometer ABL-2; Medtron, Chicago, Des Plaines, IL, U.S.A.) together with measurements of pH and pCO2. Coronary sinus venous O2 tension was measured continuously on-line with an O2 Clark type electrode (model 203B; Instech, Plymouth Meeting, PA, U.S.A.). Percentage O2 extraction (%O2E) was calculated [100 × (PO2a − PO2v/PO2a)] to assess direct vasodilatory responses apart from those due to an autoregulatory response (e.g., a decrease in CF secondary to decreased contractility). Myocardial O2 consumption (MVO2) was calculated as CF(PO2a − PO2v) times O2 solubility = 24 μl/ml saline at 760 mm Hg and 37°C.
Protocol. In nine hearts, NP and ZAP were infused individually at nine increasing log doses from 10−8 to 10−4M for 2 min each to determine the concentrations that increased CF and coronary sinus pO2 to 50% of the maximal response to 10−4M NP and ZAP (IC50 values); the drug infused first was randomized. IC50 values were estimated for each experiment by best fitting a sigmoid dose-response curve using a four-parameter logistic equation, the Hill equation (Prism; GraphPad Software, San Diego, CA, U.S.A.). In a following series of 15 hearts, NP and ZAP were infused individually and together at three concentrations of NP and ZAP above and below their IC50 values. For each drug, these concentrations were designated low, medium, and high.
Isolated aortic ring studies
Aortic rings (n = 112) were isolated from 27 guinea pig hearts prepared as before and were mounted and prepared for isometric recording in a physiologic saline solution (PSS) as reported previously (13). Care was taken not to damage the endothelium. A 6-mm portion of the aorta from the coronary ostia to the distal perfusion cannula was removed and placed in cold, oxygenated, PSS of the following composition (in 10−3M): NaCl, 119; KCl, 4.7; MgSO4, 1.17; CaCl2, 1.6; NaHCO3, 27.8; NaH2PO4, 1.18; EDTA, 0.026; glucose, 5.5; and HEPES [4(2-hydroxy-ethyl)-1-piperazineethane-sulfonic acid], 5. The aortae were divided into six to eight rings of 3 mm in length and mounted on tungsten triangles, the lowermost of which was fixed, whereas the other was attached to a force-displacement transducer (Grass model 103; Astro-Med, Inc., West Warrick, RI, U.S.A.) and suspended in jacketed temperature-controlled (37°C) tissue baths constantly aerated with a mixture of O2 and CO2 and adjusted to produce a pH of 7.4. Each ring was gradually stretched to a predetermined optimal resting tension and allowed to equilibrate for ≥90 min. Optimal resting tension was calculated in preliminary studies by measuring the contractile response to 4 × 10−2M KCl. Resting tensions were ∼2 g. The transducer signal was processed with an analog-to-digital converter (MacLab/8; AD Instruments, Mountain View, CA, U.S.A.) and recorded on computer diskettes for later detailed analysis.
Functional integrity of each vessel ring was first confirmed by the contractile response to addition of 4 × 10−2M KCl to the bath medium. Washout of the high-KCl solution with normal PSS verified return of tension toward initial levels. After a stable reproducible response to KCl was obtained (1-2 h), vessels were preconstricted continuously with 10−6M norepinephrine, a concentration found in trials to cause an ∼50% increase in tension in rings maximally relaxed by 10−4M papaverine. Parallel controls from the same vessel to the incremental addition of KCl were conducted to ensure that responses to KCl did not vary with time. Only norepinephrine-preconstricted rings displaying a >70% relaxation to 10−6M acetylcholine were used.
Relaxant responses to cumulative additions of 10−9-10−5M NP (n = 35) and 10−8 to 3 × 10−5M ZAP (n = 35) were measured. Next interactive effects of NP and ZAP (n = 42) were examined by exposing rings to six concentrations of ZAP (10−8 to 3 × 10−5M) at each of six concentrations of NP (10−9-10−5M). After completion of each experiment in the protocol, 10−4M papaverine was added to determine maximal vasodilation of each vessel ring. Results are presented as the observed relaxation induced by NP and ZAP expressed as a percentage of maximal relaxation induced by papaverine for each ring. IC50 values for NP and ZAP were determined at half the maximal relaxant response to these drugs using the Hill equation as earlier. Responses of all rings in each group belonging to each individual animal were averaged and reported as an n value of 1.
Measure of cyclic GMP
cGMP was measured in norepinephrine-preconstricted aortic ring segments (n = 40) exposed to 3 × 10−6M ZAP, 3 × 10−7M NP, both concentrations of NP and ZAP, or no drug (control) for 10 min. At that time, the aortic rings were immediately immersed in liquid N2. The rings were individually homogenized in 6% trichloroacetic acid and centrifuged at 3,000 g for 15 min. Trichloroacetic acid was extracted from the supernatant with water-saturated ether. cGMP content was then measured in aliquots of the supernatant by radioimmunoassay (code RPA 525; Amersham International, plc, Amersham, U.K.). Aortic ring protein concentration was determined by the Lowry method.
All data are expressed as mean ± standard error of the mean (SEM). Mean values were considered significant at p < 0.05. The following comparisons were made for all variables measured: Individual NP and ZAP effects (*) compared with the initial drug-free control (Control); and NP + ZAP versus ZAP alone (†). Statistical inferences were determined by two-way analysis of variance; if a significant interaction was determined, the differences between individual means were assessed by least significant difference (LSD) tests (Super Anova; Abacus Concepts, Inc. Berkeley, CA, U.S.A.). For the Hill equation, the significance of a shift in the slope at the half-maximal response (IC50) for NP versus ZAP was determined by two-way analysis of variance.
Isolated heart studies
Low, medium, and high concentrations of ZAP, 10−6, 5 × 10−6, 10−5M, and NP, 10−7, 5 × 10−7, 10−6M, respectively, given singly or combined, did not alter LV pressure (control, 105 ± 3 mm Hg; high ZAP + NP, 98 ± 5 mm Hg) or AV conduction time (control, 58 ± 2 ms; high ZAP + NP, 59 ± 2 ms). Figure 1 shows that only high NP given alone increased heart rate by 11%. Figure 2 shows that each concentration of ZAP and NP increased CF, and that ZAP + NP nonsignificantly increased CF more than either drug alone. Percentage increases in CF from control (* vs. control) for low, medium, and high concentrations, respectively, were 9.1 ± 1.2*, 9.6 ± 1.0*, 12.2 ± 2.2*% for ZAP; 9.2 ± 1.1*, 10.2 ± 1.4*, 13.7 ± 2.5*% for NP; and 15.4 ± 1.9*†, 21.2 ± 2.3*†, 23.6 ± 3.7*†% for ZAP + NP († vs. ZAP alone). CF was increased maximally by 23 ± 2*% with 10−4M NP, by 26 ± 2*% with 10−4M ZAP, and by 98 ± 4*% with a bolus injection of adenosine to elicit maximal CF in arrested hearts.
Figure 3 shows that coronary sinus pO2 (pO2cs) was increased by ZAP and NP alone, but significantly more so by ZAP + NP. Percentage increases in pO2cs from control for low, medium, and high concentrations, respectively, were 20 ± 4*, 47 ± 4*, 73 ± 5*% for ZAP; 23 ± 2*, 47 ± 4*, 64 ± 5*% for NP; and 44 ± 4*†, 78 ± 6*†, 108 ± 6*†% for ZAP + NP. Figure 4 shows that percentage O2 extraction (%O2E) was decreased by ZAP and NP alone, but significantly more so by ZAP + NP. Percentage decreases in %O2E from control for low, medium, and high concentrations, respectively, were 7 ± 2*, 17 ± 3*, 26 ± 4*% for ZAP; 9 ± 2*, 18 ± 3*, 23 ± 4*% for NP; and 17 ± 3*†, 28 ± 4*†, 40 ± 4*†% for ZAP + NP. For data not displayed, myocardial O2 consumption (MVO2) for control, low, medium, and high concentrations, respectively, were 69.3 ± 2.9, 71.4 ± 3.0, 64.0 ± 4.1, and 62.0 ± 4.2 μl/g/min for ZAP; 72.2 ± 3.2, 70.7 ± 3.4, 65.0 ± 4.1, and 63.8 ± 4.2 μl/g/min for NP; and 71.8 ± 2.8, 63.8 ± 3.5, 57.3 ± 4.0*, and 51.0 ± 4.3*† μl/g/min for ZAP + NP.
Figure 5 displays the increases in CF and pO2cs at increasing log concentration of ZAP by 50% (IC50). The IC50 values for ZAP were CF, 3.6 ± 0.1 × 10−6M, and pO2cs 3.7 ± 0.1 × 10−6M. For data not displayed, the IC50 values for NP were CF, 0.76 ± 0.09 × 10−6M, and pO2cs, 0.73 ± 0.18 × 10−6M. The isolated heart data indicate that NP and ZAP were similarly efficacious but that NP was ∼4.5 times more potent than ZAP in increasing CF or pO2cs; only the highest concentrations of ZAP + NP reduced MVO2.
Isolated aortic ring studies
Figure 6 shows the individual effects of ZAP and NP on relaxation of preconstricted aortic rings. The IC50 was 6.05 ± 0.08 × 10−6M for ZAP and 0.38 ± 0.05 × 10−6M for NP, so that ZAP was ∼15 times less potent than NP in relaxing aortic rings. Figure 7 displays the effect of NP given with ZAP on producing relaxation. The IC50 for NP is shown as a second-order polynomial function of logZAP. If the relation is made linear, the relation was logarithmic, so that IC50 [NP] (in nM) = −(60log[ZAP]) − 121; R2 = 0.97. Thus,the NP IC50 was 0.36 × 10−6M at 10−8M ZAP and 0.18 × 10−6M at 10−5M ZAP. Figure 8 shows that aortic cGMP (in fmol/mg protein) was 35 ± 29 for controls, 269 ± 39 for 3 × 10−6M ZAP, 328 ± 47 for 3 × 10−7M NP, and 523 ± 78 for ZAP + NP.
The major findings are that (a) ZAP is as effective as but less potent than NP in causing coronary vasodilation in isolated, crystalloid perfused guinea pig hearts and in preconstricted aortic rings; (b) submaximal concentrations of ZAP given with NP increase CF and pO2cs and decrease %O2E more than either drug alone; (c) except for a slight increase in heart rate by NP, and a small decrease in MVO2 with high ZAP + NP, neither ZAP nor NP have significant cardiac functional or metabolic effects; and (d) ZAP given with NP logarithmically enhances vasorelaxation of preconstricted aortic rings by increasing cGMP levels.
In intact hearts, the CF and pO2cs IC50 values were similar for NP and for ZAP, but ZAP was one fifth as potent as NP at the IC50 values. Both drugs alone increased CF by ∼25%. High concentrations of NP or ZAP equivalently inhibited CF autoregulation, as observed by their ∼68% increases in pO2cs, and 24% decreases in %O2E. Moreover, when given together, the increase in pO2cs, and the decrease in %O2E, was enhanced to 108 and 40%, respectively, of the drug-free controls. Only the higher concentrations of ZAP + NP reduced MVO2, likely because pO2cs approached the maximum. As high as 10−4M ZAP had no effect on contractile function, AV conduction time, or heart rate. Only 10−4M NP produced a slight sinus tachycardia without altering contractile function or AV conduction time. We reported previously that ≤10−5 ZAP did not alter coronary sinus release of NO in intact guinea pig hearts, whereas 10−4M NP increased coronary effluent NO concentration by 22 ± 4 nM(11).
In isolated aortic rings, the potency ratio of NP to ZAP was ∼15:1 or about 3 times that in isolated hearts. It could not be ascertained why the intact coronary vasculature response was relatively greater to ZAP than it was in preconstricted aortic rings. A major difference between these models is that the intact coronary circulation is a parallel system of vessels of many dimensions with varying degrees of smooth muscle, of which the precapillary sphincter exerts the major resistance to flow. The aorta, in contrast, is primarily a conduit for flow. Nevertheless, ZAP and NP together promoted greater vasorelaxation than did either alone. In the aortic studies, each 10-fold increase in ZAP caused subsequently much larger decreases in the NP IC50. A ZAP concentration of ∼5 × 10−6M reduced the IC50 for NP by half. Our study suggests that an additive or less than additive vasodilatory effect can be attained by a nitrovasodilator in the presence of a cGMP-dependent PDE inhibitor in these two models. The measure of cGMP in submaximally relaxed aortic rings confirms that both stimulated production of cGMP by guanylyl cyclase and inhibited degradation of cGMP by PDE V can contribute to aortic vasorelaxation.
The vascular effects of NP and ZAP have been extensively studied in isolated vessels and in intact animals; however, ZAP alone or with NP has not been examined in intact hearts devoid of nervous and hormonal influence. We have demonstrated that vasorelaxation can be elicited by endothelium-dependent and -independent drugs in these models (11-13). ZAP may induce endothelium-dependent relaxation, in part, by enhancing effects of NO release (5). In this study we assessed the single and combined effects of these cGMP-dependent vasodilators on coronary vascular tone in the intact heart and demonstrated the effectiveness of PDE V inhibition in the intact heart both in the absence and presence of stimulated cGMP synthesis.
The maximal capacity for cGMP synthesis by guanylyl cyclase is believed to be less than that for degradation by PDEs in intact cells (4). Of five major PDE isoenzymes identified in aortic vascular smooth muscle, one (type V) is cGMP specific (14). cGMP is almost exclusively hydrolyzed by PDE V in rat aortic smooth muscle cells, so cGMP levels are largely controlled by PDE V (15). Because PDE activity is regulated and normally submaximal, inhibiting PDE activity increases cGMP, as we and others have shown. In isolated rat (16) and rabbit (5,16) aortae, ZAP increased cGMP, but not cAMP, and produced relaxation of the phenylephrine-constricted isolated aorta (17). ZAP has been shown to reduce blood pressure like NP in anesthetized rats, but not to produce so great a tachycardia as NP (18). ZAP pretreatment was shown to potentiate the depressor response in rats made hypertensive by renal artery ligation (19). PDE V inhibitors, like ZAP or sildenafil (Viagra), which is ∼240 times more potent than ZAP (20), may be useful in the treatment of angina, particularly if guanylyl cyclase activity is maximally stimulated by endogenous or exogenous NO, or if NO synthase activity is reduced because of injury to the coronary endothelium. Even so, PDE V isoenzymes are unlikely to be more important in regulating cGMP levels than is synthesis of cGMP. Some PDE isoenzymes hydrolyze both cAMP and cGMP, just as various substrates are catalyzed by guanylyl cyclase to produce cGMP from cGTP.
These and other studies suggest that the combined effects of increased guanylyl cyclase activation and PDE type V inhibition on increasing cGMP levels may be of therapeutic importance for enhancing vasodilation in thromboembolic or vasospastic coronary vessels. Much as PDE III (cGMP inhibited) inhibitors like milrinone, which retards cAMP degradation, have found cardiotonic and vasodilatory utility in patients with acute heart failure, PDE V inhibitors may be useful to selectively enhance smooth muscle vasodilation, without increasing cardiac myocyte Ca2+. PDE V isoenzymes are abundant in vascular smooth muscle (21) but may not exist in myocytes (22).
Elevation of coronary cGMP levels by PDE V inhibition may also confer cardioprotection. One study showed that inhibition of soluble guanylyl cyclase raised the incidence of reperfusion-induced ventricular fibrillation and that this was abolished by cotreatment with ZAP, which increased cGMP levels (23). The increase in cGMP might have attenuated reperfusion injury by improving basal CF and responses to endogenous and exogenous nitrovasodilators, which are typically reduced on reperfusion (12). Moreover, elevation of cGMP by PDE V inhibition may counteract the decrease in cGMP formation arising from reduced stimulated NO production on reperfusion after ischemia, as shown in isolated hearts (12). Thus, there appears to be a potential benefit of PDE V inhibition for producing coronary vasodilation in some disease states over the use of drugs that stimulate cGMP synthesis like endogenous NO or nitrovasodilators. However, they must be used with caution because the combined use of these drugs could lead to coronary steal in steal-prone individuals or lead to severe hypotension.
In conclusion, this study suggests that simultaneously inhibiting cGMP hydrolysis, while enhancing cGMP production by activation of guanylyl cyclase, is a useful approach to induce coronary vasodilation without directly altering cardiac mechanical or electrophysiologic function. The vasodilatory effect of drugs that inhibit cGMP-dependent phosphodiesterase may substitute for, or be enhanced by, endogenously produced NO or administered nitrovasodilators that stimulate cGMP production.
Acknowledgment: This study was supported by Anesthesiology Research Training Grant GM 08377, VA Merit Grant 8204-04P, and AHA-Wisconsin Grant HL34708. We thank J.S. Heisner, F. Moore, M. Ziebell, C.S. Knop, and B.M. Graf for their assistance in this project. Zaprinast was a gift from Rhone-Poulenc Rorer Ltd.
1. Waldman SA, Murad F. Biochemical mechanisms underlying vascular smooth muscle relaxation: the guanylate cyclase c-GMP system. J Cardiovasc Pharmacol
2. Busse R, Mulsch A, Fleming I, Hecker M. Mechanisms of nitric oxide release from the vascular endothelium. Circulation
3. Moncada S, Higgs A. The L-arginine-nitric oxide pathway. N Engl J Med
4. Sonnenburg WK, Beavo JA. Cyclic GMP
and regulation of cyclic nucleotide hydrolysis. Adv Pharmacol
5. Martin W, Furchgott RF, Villani GM, Jothianandan D. Phosphodiesterase inhibitors
induce endothelium-dependent relaxations of rat and rabbit aorta by potentiating the effects of spontaneously released endothelium-derived relaxing factor. J Pharmacol Exp Ther
6. Pagani ED, Buchholz RA, Silver PJ. Cardiovascular cyclic nucleotide phosphodiesterases and their role in regulating cardiovascular function. Basic Res Cardiol
7. Komas N. Le Bec A, Stoclet JC, Lugnier C. Cardiac cGMP-stimulated cyclic nucleotide phosphodiesterases: effects of cGMP analogues and drugs. Eur J Pharmacol
8. Polson JB, Strada SF. Cyclic nucleotide phosphodiesterase and vascular smooth muscle. Annu Rev Pharmacol Toxicol
9. Merkel LA, Rivera LM, Perrone MH, Lappe RW. In vitro and in vivo interactions of nitrovasodilators and zaprinast
, a cGMP-selective phosphodiesterase inhibitor. Eur J Pharmacol
10. Pagani ED, Vanller GS, O'Connor B, Silver PJ. Reversal of nitroglycerin tolerance in vitro by the cGMP-phosphodiesterase inhibitor zaprinast
. Eur J Pharmacol
11. Fujita S, Roerig DL, Bosnjak ZJ, Stowe DF. Effects of vasodilators and perfusion pressure on coronary flow
and simultaneous release of nitric oxide from guinea pig isolated hearts. Cardiovasc Res
12. Novalija E, Fujita S, Kampine JP, Stowe DF. Sevoflurane mimics ischemic preconditioning effects on coronary flow
and nitric oxide release in isolated hearts. Anesthesiology
13. Stowe DF, O'Brien WC, Chang D, Knop CS, Kampine JP. Reversal of endothelin induced vasoconstriction by endothelium dependent and independent vasodilators in isolated hearts and vascular rings. J Cardiovasc Pharmacol
14. Saeki T, Saito I. Isolation of cyclic nucleotide phosphodiesterase isozymes from pig aorta. Biochem Pharmacol
15. Coste J, Grondin P. Characterization of a novel potent and specific inhibitor of type V phosphodiesterase. Biochem Pharmacol
16. Lugnier C, Schoeffter P, Le Bec A, Strouthou E, Stoclet JC. Selective inhibition of cyclic nucleotide phosphodiesterases of human, bovine and rat aorta. Biochem Pharmacol
17. Weishaar RE, Kobylarz DC, Keiser J, Haleen SJ, Major TC, Rapundalo S. Subclasses of cyclic GMP
-specific phosphodiesterase and their role in regulating the effects of natriuretic factor. Hypertension
18. Ng SSW, Pang CCY. Zaprinast
, a type V phosphodiesterase inhibitor, dilates capacitance vessels in anaesthetised rats. Eur J Pharmacol
19. Anderson IK, Drew GM. Investigation of the inhibitory effect of Ng
-nitro-L-arginine methyl ester on the antihypertensive effect of the angiotensin AT1
receptor antagonist, GRI38950. Br J Pharmacol
20. Ballard SA, Gingell CJ, Tang K, Turner LA, Price ME, Naylor AM. Effects of sildenafil on the relaxation of human corpus cavernosum tissue in vitro and on the activities of cyclic nucleotide phosphodiesterase isozymes. J Urol
21. Perry MJ, Higgs GA. Chemotherapeutic potential of phosphodiesterase inhibitors
. Curr Opin Chem Biol
22. Bode DC, Kanter JR, Brunton LL. Cellular distribution of phosphodiesterase isoforms in rat cardiac tissue. Circ Res
23. Pabla R. Bland-Ward P, Moore PK, Curtis MJ. An endogenous protectant effect of cardiac cyclic GMP
against reperfusion-induced ventricular fibrillation in the rat heart. Br J Pharmacol