Acta Physiol Scand 1990, 140, 209-219

Effects of isozyme-selective phosphodiesterase inhibitors on rat aorta and human platelets: smooth muscle tone, platelet aggregation and cAMP levels S.H. S. L I N D G R E N , T. L. G. ANDERSSON, E. V I N G E and K.-E. ANDERSSON Department of Clinical Pharmacology, University Hospital, Lund, Sweden LINDGREN, S. H. S., ANDERSON, T . L. G., VINGE,E. & ANDERSON, K.-E. 1990. Effects of isozyme-selective phosphodiesterase inhibitors on rat aorta and human platelets : smooth muscle tone, platelet aggregation and cAMP levels. Acta Physiol Scand 140, 209-219. Received 3 January 1990, accepted 27 April 1990. ISSN 0001-6772. Department of Clinical Pharmacology, University Hospital, Lund, Sweden. The inhibitors of the cGMP-inhibited, low-K, cAMP phosphodiesterase-milrinone and OPC 3911-and an inhibitor of a non-cGMP-inhibited low-K, cAMP phosphodiesterase-rolipram-were used to evaluate the functional importance of the two cAMP phosphodiesterase activities in vascular smooth muscle and in platelets. Vinpocetine, an inhibitor of a calcium-calmodulin-dependent phosphodiesterase was also studied. OPC 3911 and milrinone relaxed the contracted rat aorta, inhibited ADPinduced platelet aggregation and also enhanced isoprenaline-induced relaxation as well as the antiaggregatory effects of adenosine. In platelets, OPC 3911 and milrinone increased cAMP levels, but in the rat aorta the increase was significant only for milrinone (OPC 3911 P = 0.062). In both tissues OPC 3911 and milrinone enhanced the increase in cAMP caused by activators of adenylate cyclase (isoprenaline/adenosine). Rolipram had no effects on aggregation or cAMP levels in platelets and no overadditive effects in combination with adenosine. Rolipram had little effect on relaxation and cAMP levels, did not alter isoprenaline-induced relaxation of guanfacin-contracted rat aorta, but showed synergistic effects with isoprenaline in raising cAMP levels. In PGF,,-contracted aorta rolipram enhanced relaxation caused by isoprenaline. Vinpocetine had a relaxant effect without affecting cAMP levels, but had no effect on platelets. These results support the concept that the cGMP-inhibited phosphodiesterase is an important modulator of vascular smooth muscle tone and platelet function. The role of the non-cGMP-inhibited phosphodiesterase in these tissues is less obvious. Key words : adenosine, CAMP, human platelets, isoprenaline, milrinone, OPC 39 11, phosphodiesterase inhibitors, rat aorta, roliprarn, synergism. Relaxation of vascular smooth muscle and inhibition of blood platelet aggregation can be mediated by cyclic adenosine 3’,5’-monophosphate (CAMP) generation. Breakdown of cAMP is controlled by cyclic nucleotide phosphodiesterase (PDE) activity. The existence of different isozyme families of cyclic nucleotide Correspondence : Sam Lindgren, Department of Clinical Pharmacology, University Hospital, S-22 1 85 Lund, Sweden.

PDE is now apparent, and several drugs are known to selectively inhibit individual PDE isozymes (for reviews see Beavo 1988, Silver 1989). I n earlier studies, several PDE isozymes with differences in their regulation, substrate affinities and rate of hydrolysis have been described in, for example, vascular smooth muscle and blood platelets (Hidaka & Asano 1976, Lugnier et al. 1986, Weishaar et al. 1986). I n both tissues, a low-K, CAMP-specific PDE isozyme has been

209

210

S . H . S . Lindgreri et al.

connected to a Grass polygraph. The tissues were allowed to equilibrate for a period of 0.5-1 h with repeated washings. The initial mounting tension was adjusted until a stable tension of 10 mN was maintained. -411 ring segments were initially contracted by exposure to a high- (124 mM) K' solution, prepared by substituting all NaCl in the standard Krebs solution with equimolar concentrations of KCI. The following series of experiments was performed : (1) Concentration-response curves were obtained for guanfacin in the presence of prazosin 0.1 PM or rauwolscine 1 p~ added 15 min prior to guanfacin. Vessel segments which were exposed to guanfacin in parallel served as controls. Contractions were expressed as a percentage of the mean of two reproducible contractions (variation < 1006) induced by K- 124 mM. (2) PDE inhibitors or isoprenaline were added cumulatively to aortic rings contracted with guanfacin 0.3 or guanfacin 3 p ~ . (3) When a contraction elicited by guanfacin 3 ,UM had stabilized, concentrations of the PDE inhibitors relaxing preparations by about loo/,, or their vehicles, were added. After 15 min isoprenaline or its vehicle was added cumulatively. When the concentrationresponse curve for isoprenaline was completed (20 min) the preparations were immediately immersed in liquid nitrogen, placed into test tubes containing loo;,, trichloroacetic acid, and stored at -70 "C until assayed for CAMP. (4) Milrinone 30 ,UM, OPC 391 1 30 ,UM, rolipram 30 ,mi or the respective vehicle was added to guanfacin 30 I ~ Mcontracted aorta. After 15 min a single concentration of isoprenaline 1 ,UM or vehicle was added. When maximal relaxation occurred (7 min) the tissues nere rapidly immersed in liquid nitrogen and treated as described above. (5) The same procedure was used as in expt 3, but the contraction was induced by 10 p M PGF,, in the presence of prazosin 0.1 ,UM added 15 min prior to PGF,,. CAMP was not measured. M A T E R I A L S AND 31E T €1 0 D S Control measurements of vehicle effects were performed on separate aortic rings. Relaxation was .Weuswentent id aortic tensron expressed as a percentage of maximum relaxation, i.e. Xlale Sprague-Dawley rats (150-250 g) (llollegaard back to baseline tension. Breeding Center, Denmark) were sacrificed h\- cervical dislocation. The thoracic aorta nas remo\-ed and Platelet aggregation studies cleaned of adventitial connective tissue. -4 metal probe Venous blood from healthy drug-free volunteers was n a s introduced into the lumen of the vessel, thus mixed with 1/9 volume of acid citrate-dextrose destro! ing the endothelium. Tension studies were solution, and platelet-rich plasma (PRP) was prepared performed as preyiously described (Hogestatt et al. as previously described (Anderson & Vinge 1988). 1983). Briefly, ring segments 2 mm in length were Aggregation was performed in a Chrono-Log model mounted in temperature-controlled (37 "C) organ 540 aggrometer with a Hitachi recorder. Samples baths containing 5 ml of gassed (95 O,, O,, 5 O0 CO,) (0.45 ml) of PRP were preincubated at 37 "C for about Krebs solution of the following composition (mM) : 15 min, and stirred for 1 min before addition of the NaCI 119, NaHCO, 15, KCI 4.6, CaCI, 1.5, NaH,PO, PDE inhibitors or saline. Subsequently adenosine or 1.2, XlgCI, 1.2 and glucose 11 (pH z 7.4). Isometric saline and then adenosine diphosphate (ADP) 2 ,UM tension was measured by a Grass FT03C transducer were added at 1-min intervals. For the study of the

identified. Recent]!. it was shown that the lowK,,, c.-2MP PDE from bovine aorta consists of two isozymes (Rascon r t al. 1990). One of these is a CAMP PDE nhose activity is greatly inhibited by low concentrations of c G M P (cGIPDE) and also b!- drugs such as cilostamide and milrinone (a bipyidine deri\-ative). It has been demonstrated in human and bovine platelets that the hydrolysis of cA3,lP by cGI-PDE represents more than 8O0,, of the total low-Kn, cAXlP activity (YlacPhee et a / . 1986). T h e other lowK,,, c491P PDE isozyme (RI-PDE) is not inhibited by low concentrations of c G M P , but is selectively inhibitcd by rolipram (Beavo 1988). .4 functional role for cGI-PDE has been indicated in both vascular smooth muscle and platelets (Weavo 1988). -4s to RI-PDE, rolipram has been shown to have vasorelaxant effects in n f r o (Schoeffter et a / . 1987) but little influence on platelet function (Simpson et (11. 1988). 130~.ever, there are conflicting results regarding the ability of cGI-PDE and RI-PDE inhibitors to enhance the functional and biochemical effects of- activators of adenylate cyclase (Schoeffter et ill. 1987, Tanaka et a / . 1988, Jackson et 01. 1989). In the present investigation the cGI-PDE inhibitors milrinone and OPC 3911 (a new cilostamide derivative) and the RI-PDE inhibitor rolipram were used to evaluate further the functional importance of the two c A M P PDE activities in vascular smooth muscle and blood platelets. Vinpocetine, a proposed inhibitor of a Ca'+/calmodulin-dependent PDE (Ca'+/CaM PDE), which preferentially hydrolyses c G M P (Hagiwara et a / . 1981), was included in the study for comparative purposes.

PDE inhibitors, rat aorta and platelets combined effects of adenosine and the PDE inhibitors, adenosine 1 ,UM and subthreshold concentrations of the PDE inhibitors were used. Aggregation was measured as the maximum slope (V,,,) of the aggregation curve, and expressed as percentage of aggregation induced by ADP 2 ,UM (control). All tests were performed in duplicate for each drug concentration. Saline controls were run after every two or three samples with any drug, and the sequence of concentrations was changed between the runs. Measurement of CAMP

cAMP levels in rat aorta. Determination of cAMP content was performed in tissues used for measurement of tension, according to the protocol outlined above (expts 3 and 4). In a separate series of experiments aortic rings were only incubated and not suspended in the tissue baths, but the conditions were otherwise identical. Following exposure to guanfacin 3 ,UM for 20 min, the PDE inhibitors were added in concentrations that in functional studies caused a relaxation of 10% (expt 3), and after an additional 15min period a single concentration of isoprenaline 1 ,ILM or vehicle was added. After another 4 min of incubation the tissues were immersed in liquid nitrogen, and handled as described in expt 3. Frozen aortic tissue (x 3 mg) was homogenized at 4 "C in 1.2 ml of 10% trichloroacetic acid (TCA) using a glass-glass homogenizer, and centrifuged at 2000 g (4 "C) for 15 min. The pellets were reconstituted in 0.5 M NaOH and tested for protein content (Bradford 1976). The supernatants were extracted five times with 5 ml of water-saturated diethyl ether. The aqueous phase was evaporated and the residue stored at - 20 "C. Residues were dissolved in 0.05 M sodium acetate, pH 6.2, and the amount of cAMP was quantitated using [1251]~AMPRIA kits (RIANEN, Du Pont) according to kit instructions. Cyclic nucleotides were acetylated with acetic anhydride to increase the sensitivity of the assay. The final values of tissue cAMP were corrected for recoveries ( c 85 %) determined by addition of trace amounts of [3H]cAMP to tissue homogenates. cAMP levels in platelets. PRP was centrifuged for 10 min at 1000 g, and the pellet resuspended in a buffer containing (mM) Hepes 10, NaCl 145, KCI 5, MgSO, I , Na,HPO, 0.5 and glucose 6, pH 7.4, at 37 "C. The platelet count was adjusted to approximately 200 x lo9 I-'. Samples (0.5 ml) of washed platelets were incubated at 37 "C for 30 min before adding OPC 391 1, milrinone, rolipram or saline. After 1 min adenosine or saline was added, and after another minute the samples were quenched with icecold 10% TCA, vortexed and put on ice. [3H]cAMP (50 ,d)was added to each sample, before centrifugation at 2500 g (4"C) for 10 min, and the supernatants were then treated as described above. All values were corrected for recoveries (mean 747,). Each con10

211

centration and combination of drugs was tested in duplicate, except saline controls, which were tested in quadruplicate. Drugs Drugs used were OPC 391 1 (Otsuka Pharmaceuticals), milrinone (Sterling-Winthrop Research Institute), rolipram (Schering), vinpocetine (Gedeon Richter), guanfacin hydrochloride (Sandoz), isoprenaline hydrochloride (Sigma), PGF,, (Amoglandin, Astra), prazosin hydrochloride (Pfizer), rauwolscine hydrochloride (Carl Roth), adenosine and ADP (Sigma). Dilution of stock solutions were made with sodium chloride 154 mM with ascorbic acid (1 mM). Statistics Data are given as mean SEM ;n denotes the number of aortic rings examined (from a minimum of four rats) or the number of volunteers donating blood. The EC,, values, defined as the concentration giving 50% of maximal response obtained, were determined graphically and expressed as the negative logarithm (pEC,,). Responses to relaxant/antiaggregatory agonists were compared by the use of geometric mean EC,, values and maximum responses within the concentration range used. The cAMP levels were evaluated comparing changes in absolute values. Student's t-test (two-tailed) was used for comparison of means (unpaired for aorta, paired for platelets). The initial small relaxation caused by the PDE inhibitor was corrected for when the relaxant effect of isoprenaline added cumulatively was calculated. Dunnet's t-test was then used to compare the responses of isoprenaline in the presence or absence of a PDE inhibitor. A P-value less than 0.05 was considered significant.

RESULTS Functional studies: rat aorta Antagonism of guanfacin eflects. The response to guanfacin (10 nm-100 ,UM) was almost abolished in the presence of prazosin 0.1 ,UM, while in the presence of rauwolscine 1 ,UM there was an approximately fivefold shift to the right of the concentration-response curve for guanfacin. T h e EC,, value for guanfacin without pretreatment was 0.24 ,UM (pEC,, = 6.6 k 0.04) and in the presence of rauwolscine 1.3 ,UM (pEC,, = 5.9 f0.1). Concentration-response

relationships.

Iso-

prenaline and the PDE inhibitors, with the exception of rolipram, had significantly reduced capacity to relax the rat aorta when the concentration of guanfacin was increased from ACT 140

212

S . H. S. Lindgren et al.

Table 1. Effects of cumulative addition of the PDE inhibitors (0.01-30 PM) and isoprenaline (0.01-100p~) on guanfacin 0.3 ,UM and 3 PM contracted rat aortic rings

EC," Guanfacin 0.3 ,UM OPC 3911 (n = 7) Milrinone (n = 7) Rolipram (n = 6) Vinpocetine (n = 7) Isoprenaline (n = 8) Guanfacin 3 PM OPC 3911 (n = 8) Milrinone (n = 7) Rolipram (n = 8) Vinpocetine (n = 7) Isoprenaline (n = 8)

+

Date are means SEM. *; P < 0.05, ** P < 0.01,

(M)

4.0 x lo-' 1.0 x 10-6 1.4 x 4.4 x 10-6 7.6 x 4.9 x 8.3 x 8.5 x 9.8 x 7.8 x

10-6

lo-' 10-6 lo-'

PEC,,

Maximal response (yo)

6.4 k 0.1 6.0k0.2 5.8k0.4 5.4kO.1 7.2 & 0.1

97+2 97+ 1 33+ 12 7654 82f9

5.3 k0.2*** 5.1 &O.l*** 6.1 k 0.2

67 6** 84 & 2*** 12+4 44+ lo* 47+ 13*

5.0k0.1f* 6.1 f0.2***

+

*** P < 0.001 t's corresponding value at guanfacin 0.3 PM.

0-0

25 -

E

Y

C

8 2

50-

-a, r

75 -

I

100

1

*I

,

Pretreatment

I

10-8

I

10-7

10-6

10-5

10-4

Isoprenaline conc ( M )

Fig. 1. Cumulative isoprenaline concentration-response curves in rat aortic rings contracted with 3 p~ guanfacin. The tissues were pretreated with OPC 3911 0.3 ,UM ( O ) ,milrinone 1 p~ (m), rolipram 3 0 p ~ (O), vinpocetine 5 p~ (A, n = 11) or vehicle (a).Data are means$.SEM; n = 12 if not stated otherwise. * P < 0.05, ** P < 0.01 vs vehicle-treated tissues.

PDE inhibitors, rat aorta and platelets

75

I

213

*:*

*** T

-

-8

50

C

. c

1 a

25

0

I

OPC

Is0

30pM 1pM

Mil Is0 30pM 1pM

oyc Is0

Mil ls’o

Rol Is0 30pM 1pM

Rol

lco

Fig. 2. Relaxation of guanfacin 30 ,LAMcontracted rat aorta. The effects of OPC 391 1 30 p~ (n = 9), m i h o n e 30 ,UM (n = lo), rolipram 30 p~ (n = 7), isoprenaline 1 ,EM (n = 7-10) and the combination of each PDE-inhibitor and isoprenaline were studied in matched aortic rings. Data are means & SEM. *** P < 0.01 compared with the values calculated by assuming additivity of the

drugs.

1 **

1 **

lsoprenaline conc ( M )

Fig. 3. Cumulative isoprenaline concentrationresponse curves in rat aortic rings contracted with 10 p~ PGF,,. The tissues were pretreated with OPC 3911 0.3 ,UM (O), milrinone 1 p~ (H),rolipram 30 ,UM (O), vinpocetine 5 ,UM (A, n = 7) or vehicle ( 0 ) .Data are meansf SEM; n = 8 if not stated otherwise. * P < 0.05, **P < 0.01 vs vehicle-treated

tissues.

0.3 to 3 p (Table ~ 1). T h e effects of rolipram were modest at both concentrations of guanfacin. Combined effects of isoprenaline and PDE inhibitors. I n tissues contracted with guanfacin 3 p~ the relaxation caused by the cumulative addition of isoprenaline was significantly enhanced by pretreatment with OPC 3911 0.3 ,UM or milrinone 1 ,UM (Fig. 1). Pretreatment by either rolipram 30 ,UM or vinpocetine 5 ,UM did not significantly alter the relaxation obtained with isoprenaline (Fig. 1). Similar results were found when guanfacin 30 ,UM was used to maximally contract the aortic rings. At this concentration of guanfacin the relaxant effects of OPC 391 1 and milrinone were reduced while the effects of rolipram were virtually unchanged. When OPC 391 1 30 ,UM or milrinone 30 ,UM was used as pretreatment, the relaxation caused by isoprenaline 1 ,UM was significantly enhanced (Fig. 2). I n the presence of rolipram 30 p ~the , response to isoprenaline 1 ,UM was not significantly affected (Fig. 2). T h e effects of vinpocetine were not studied. With PGF,, as the contractile agonist, pretreatment with OPC 3911 0.3 p ~ milrinone , 1 ,UM or rolipram 30 p~ significantly enhanced the isoprenaline-induced relaxation (Fig. 3). Vinpocetine did not affect the actions of isoprenaline (Fig. 3). 10-2

214

S. H. 5'. Lindgren et al.

Plii trlrt, Concetztrution-2rihihrtian curres. Figure 1 shohs

the effect of increasing concentrations of the

drugs investigated. OPC 391 1 and milrinone were both potent inhibitors of ADP-induced aggregation with EC,, values of 1.6 /AM (pEC,, = 5.8+0.05) and 14 ,UM (pEC,, = 4.8i0.04) respectively. Neither rolipram nor vinpocetine ( n = 3, data not shown) had any significant effects in the concentration ranged examined (0.1-100 ,uM). Adenosine inhibited, but did not abolish, the ADP-induced aggregation (EC,, = 1.2puhi, pEC,, = 5.9i0.09). Combined efeects of adenosine and PDE anhibitors. The effects of adenosine, PDE inhibitors and the combination of the drugs on ADPinduced platelet aggregation are shown in Fig. 5. At the concentrations chosen, the PDE inhibitors alone had no effects on aggregation. Both OPC 3911 0.1 ,L'M and milrinone 1 ,/AM significantly enhanced the inhibitory effect of adenosine, whereas rolipram and vinpocetine ( n = 3, data not shown) 10 p~ did not. ,I.leusureinrnts of c-4MP content

Concentration ( M )

Fig. 4. Inhibition of ADP-induced ( 2 p h i ) plateler aggregation b> OPC 391 1 (O), milrinone (m), rolipram ( 0 ,I I = 5 ) and adenosine (-&). Data are means & SEM ; tt = 6 if not stated otherwise.

R a t riorta. The CAMP content in aortic rings contracted with guanfacin 3 ,UM was 9.5 It 1.0 pmol mg-' protein, corresponding to 105O, of tissues not treated with any drug (no significant difference). OPC 391 1 (0.3 /AM), milrinone (1 p ~ ) ,rolipram (30 p ~ or ) vinpocetine ( 5 p h i ) caused 109, relaxation (see Materials and Methods, expt 3 ) , but did not

60 -

r

*-*

I

8

I '

O k Ado 0.lw 1&4

fi 1

Y C

Mil

Ado

t$l

Ado

1pM

1&4

Ado

RoI

Ado

Rol

10pM 1pM

Ado

+

Fig. 5. Inhibition of ADP-induced (2 pLi) platelet aggregation by OPC 3911 0.1 p u ~ milrinone , I ,//.at and rofiprarn I0 p i (n = 5 ) alone and in combination with adenosine 1 p ~ Data . are means & SEM ; n = 6 if not stated otherwise. ** P < 0.0 1 compared with the values calculated by assuming additivitj-of the drugs.

PDE inhibitors, rat aorta and platelets

215

300

-8 250 --$200 >

-

4

9

150

c .-

a,

3 100 z

C

50

0 I

Mil 1pM

Is0

Mil

1pM o: I

Rol

Is0

30pM 1pM

Rol &o

Fig. 6. Changes in cAMP levels in rat aortic tissue incubated with guanfacin 3 p ~The . effects of OPC 391 1 0.3 ,uM (n = 7), mihinone 1 p~ (n = 7), rolipram 30 pM (n = 6), isoprenahe 1 PM (n = 6 7 ) and the combination of each PDE inhibitor and isoprenaline were studied in matched aortic rings. The cAMP content in controls (only guanfacin-treated) for OPC 391 I and milrinone was 7.64+0.51 and for rolipram 11.74f 1.39 pmol mg-' protein. Data are expressed as percentage change relative to controls, means fSEM. t P < 0.05, tt P < 0.01 vs controls. * P < 0.05, ** P < 0.01 compared with the values calculated by assuming additivity of the drugs. significantly alter the cAMP levels (data not (30 p ~ on) CAMP levels in guanfacin- (30 p ~ ) shown). After the cumulative addition of iso- contracted rings of aortic smooth muscle (expt 4) prenaline alone, the cAMP content was sig- are demonstrated in Fig. 7. OPC 3911 and nificantly increased ( X 1.9-fold). Of the com- milrinone each appeared to increase the cAMP bined treatments, only rolipram and isoprenaline levels, but it was statistically significant only for caused an increase in cAMP levels to milrinone (P = 0.062 for OPC 3911). No such 34 f3.5 pmol mg-' protein, which was sig- trend was observed for rolipram. In the presence nificantly higher than would have been expected of either one of the PDE inhibitors the if the increases obtained separately with rolipram isoprenaline-elicited increases in cAMP levels and isoprenaline were additive (P < 0.01). were significantly above those predicted by In aortic smooth muscle rings incubated with additivity (Fig. 7). The changes over time for guanfacin 3 p ~the, cAMP levels were approxi- isoprenaline-stimulated increases in cAMP levels mately l l O ~ oof those in untreated tissues (not were examined at 1 min (n = 4), 4 min (n = 4) significantly different). In the presence of OPC and 7 min (n = 5). The maximum increase in 3911 (0.3 p ~ ) milrinone , (1 p ~ or) rolipram cAMP levels was seen at 1 rnin (103 yo);at 4 min (30 p ~ )the , increases in cAMP levels caused by the increase was 84% and at 7 rnin 44%. Platelets. The basal level of cAMP in unisoprenaline were above those predicted assuming additivity of the drugs (Fig. 6). Vinpo- stimulated platelets was 4.0k0.13 pmol per 10' cetine at a low (5 p ~ or) at the maximal platelets (n = 5). Adenosine (10 /AM) caused a concentration (30 p ~did) not increase the CAMP 30% increase in cAMP to a level of 5.2f0.11 ) an increase levels, and did not alter the increase in CAMP (P < 0.01). OPC 391 1 (30 p ~caused content caused by isoprenaline alone (n = 4, (245%) in the levels of cAMP to 14+ 1.1 (P < 0.001). After incubation with milrinone data not shown). The effects of single concentrations of OPC (30 p ~the) cAMP levels were elevated (163 yo)to 391 1 (30 p ~ )milrinone , (30 p ~ and ) rolipram 10+0.65 (P < 0.001). The cAMP content after

216

S. H . S. Lindgren et al.

**,

3CO}

OPC

IS0

30pM 1pM

o_pc Is0

Mil

Mil + Is0

Is0

30pM 1pM

Is0

Rol

30pM 1pM

Is0

Rol

+

Fig. 7. Changes in CAMP le\-els in guanfacin 30 //M contracted rat aorta. T h e effects of OPC 391 1 30 ,UM ( n = 9), milrinone 30 /LM(n = lo), rolipram 30 / l M (n = 7 ) , isoprenaline 1 ,UM (n = 7-10) and the combination of each PDE inhibitor and isoprenaline were studied in matched aortic rings. 'The CAMP content in controls (only guanfacin-treated) was for OPC 3911 8.48k0.60; for milrinone 8.26-1:0.68; and for rolipram 9.6220.83 pmol of CAMP per mg protein. Data are expressed as percentage change relative to controls, means? SEM. tt P < 0.01, ttt P < 0.001 us controls. **P < 0.01 compared with the values calculated by assuming additivity of the drugs.

I

*-*

r

**

t

nIti Ad0

OPC

Mi

Ado

Mil

Rol

0,lpM 1pM

A+do

1pM

1pM

A'do

10pM 1pM

OPC

Ado

Rol A+do

Fig. 8. Increases in CAMP levels in platelets by OPC 3911 0.1 ,UM,milrinone 1 ,/AM and rolipram 10 ,uM alone and in combination with adenosine 1 ~ U MData . are expressed as percentage change relative to the basal level 4.0 pmol per 10' platelets. Data are meanskSEM, n = 5 . t P < 0.05, tt P < 0.01 z's basal level. **P< 0.01 compared with the values calculated by assuming additivity of the drugs.

addition of rolipram (30 ,UM) was 4.0f0.19, which n a s not significantly different from the basal le\.el. I n Fig. 8 the increases in cAhlP levels by

adenosine 1 ,BM alone a n d in combination with low concentrations of the PDE inhibitors are shown. Adenosine, OPC 3911 0.1 ,BM a n d milrinone I ,BM b u t not rolipram 10 f t caused ~ ~

PDE inhibitors, rat aorta and platelets statistically significant rises in cAMP levels. For the combination of adenosine with either OPC 391 1 or milrinone the increase was significantly greater than predicted by additivity.

DISCUSSION The data presented here give considerable support to the view that inhibition of cGI-PDE is linked to vasorelaxation and inhibition of platelet aggregation. The cGI-PDE inhibitors OPC 3911 and milrinone alone relaxed the precontracted rat aorta, and both enhanced the isoprenaline-induced relaxation. They also inhibited platelet aggregation and enhanced the antiaggregatory effects of adenosine. Rolipram, an inhibitor of the RI-PDE isozyme, had little relaxant effect, and no effects on platelets (see Simpson et al. 1988). The previously shown relaxant effects of rolipram on serotonincontracted rat aorta (Schoeffter et al. 1987) were observed at high concentrations of the drug (50% inhibition of contraction at 150 yM). Functional effects at these concentrations may be the result of non-selective PDE inhibition (see Beavo 1988) or a non-specific action. The inhibitor of the Ca2+/CaM PDE, vinpocetine, had a distinct vasorelaxant effect. As expected (see Hagiwara et al. 1984, Souness et al. 1989) vinpocetine alone did not affect cAMP levels and the combination with isoprenaline did not produce any overadditive effects on either relaxation or cAMP levels. In agreement with the results of Hidaka & Endo (1984), who reported that large amounts of the Ca2+/CaM PDE could be demonstrated in vascular tissue but not in platelets, vinpocetine had no effects on platelets.

Methodological considerations The main effect of adenosine on human platelets, within the concentration range used, is an activation of adenylate cyclase via the A,receptor. At higher concentrations adenosine has intracellular effects by direct inhibition of adenylate cyclase (Haslam & Cusack 1981). This may explain the flat concentration-response curve for adenosine. Guanfacin is considered to be predominantly an a,-receptor agonist (Dausse et al. 1983). In

217

the rat aorta, however, the contractions caused by guanfacin were effectively antagonized by prazosin, believed to be a selective blocker of aZ1receptors. Others have reported similar findings when studying the effects of the a,-receptor agonist clonidine and prazosin on rat aorta (Randriantsoa et al. 1981). It has been suggested that prazosin may block a subtype of the a2receptor (Young et al. 1989).

cAMP levels and functional effects In platelets, the effects of OPC 3911, milrinone and rolipram on cAMP levels correlated well with their inhibitory effects on aggregation. However, in the rat aorta, low concentrations of the PDE inhibitors which caused 10% relaxation did not increase cAMP levels. This is in agreement with previous observations in vascular and airway smooth muscle (Bryson & Rodger 1987, Silver et al. 1988). T h e effects of high concentrations of PDE inhibitors on cAMP content were investigated using a supramaximal concentration of guanfacin. Under these conditions the cGI-PDE inhibitors had reduced relaxant effects. OPC 3911 and milrinone both increased cAMP levels, but this was statistically significant only for milrinone. It cannot be excluded that high concentrations of cGI-PDE inhibitors may also inhibit other PDE isozymes (Harrison et al. 1986). Rolipram alone did not raise cAMP levels at any concentration studied. Preliminary results indicate that RI-PDE represents a smaller fraction of the PDE isozymes hydrolysing cAMP in the rat aorta (Lindgren et al. unpublished data; see also Souness et al. 1989). T h e discrepancies between cAMP levels and functional effects of. the cGI-PDE inhibitors in vascular smooth muscle may have several explanations, including experimental and technical inadequacies (see also Silver et al. 1988). A possibility could be that the contractile agonist decreases the basal adenylate cyclase activity, which would diminish the effects of the cGIPDE inhibitors. However, in this study the cAMP content in the rat aorta was not altered by guanfacin treatment (3 PM). Of course, our results as well as the findings by others may be interpreted as if the cGI-PDE inhibitors, besides inducing accumulation of CAMP, have additional mechanisms of action, a t least in smooth muscle.

218

S . H . S. Lindgren et nl.

rat aorta. Schoeffter et al. (1987) found that a moderate concentration of rolipram enhanced both the CAMP-increasing and the relaxing I n platelets, u-e found that under similar effects of isoprenaline when rat aorta was experimental conditions the cGI-PDE inhibitors contracted with PGF,,. In our study, the series OPC 3911 and milrinone clearly enhanced the of experiments performed on PGF,,-contracted effects of adenosine on both aggregation and aorta showed that OPC 3911, milrinone and CAMP levels, whereas inhibition of RI-PDE by rolipram all enhanced the relaxant effects of rolipram did not affect the actions of adenosine. isoprenaline (CAMP levels not measured). T h e Earlier studies have s h o m synergism between apparent discrepancy in the results obtained other selective P D E inhibitors and P G E , or with guanfacin and PGF,, may be explained if P G I , on increases in CAMP levels (Hidaka et a/. there exist different compartments of c A M P in 1979, Tang & Frojmovic 1980, Jackson et 01. rat aortic smooth muscle (see Rubanyi et al. 1989). T a n g & Frojmovic (1980) also found 1986, 1-egesna & Diamond 1986). I f guanfacin synergistic effects on inhibition of aggregation, and PGF,, induce contractions through different but used different drug concentrations in bio- intracellular pathways, cellular compartments of c A M P may affect these contractions dissimilarly. chemical and functional studies. I n summary, our functional and biochemical In the rat aorta, the relationship between P D E inhibition and enhancement of the effects of data support the concept that cGI-PDE is an isoprenaline was more complex. Low con- important modulator of vascular smooth muscle centrations of the cGI-PDE inhibitors enhanced tone and platelet function. T h e possible reguthe relaxation but not the increases in c'4MP latory role of RI-PDE is less obvious. levels induced by isoprenaline added cumulatiyely. This finding led us to question the This studj- was supported by grants from the Medical experimental procedure and whether the studied Faculty at the Lnivetsity of Lund and the Swedish concentrations of drugs were optimal. Another Society of Medicine. The authors thank Professor protocol was designed in which aortic rings were Vincent Manganiello for the gift of OPC 3911 and Professor Gyorgy Fekete for providing vinpocetine. incubated only with the drugs (no tension Lye also thank Hans-Inge Bengtsson for technical applied), using the same low concentrations of assistance with the CAMP assay. the P D E inhibitors combined with a single concentration of isoprenaline. I n these experiments cGI-PDE inhibition enhanced the R E F E R E N C E S effects of isoprenaline on c A M P levels. T h e next step was to evaluate the ability of high con- .~NDERSSON, T.L.G. & VINGE,E. 1988. Effects of ouabain on "Rb-uptake, 3H-5-HT-uptake and centrations of the cGI-PDE inhibitors to increase aggregation by 5-HT and ADP in human platelets. both the functional and biochemical effects of a PhurniaroL Toxicol 62, 172-176. single concentration of isoprenaline. I n this REAVO, J.A. 1988. Multiple isozymes of cyclic experimental model, cGI-PDE inhibition ennucleotide phosphodiesterase. In : P. Greengard & hanced both the relaxation and the increases in G..l. Robinson (eds.) ,4dvances in Second Messenger CAMP levels caused by isoprenaline. O u r results and Phosphoprotein Research, vol. 22, pp. 1-38. agree with those of Tanaka et al. (1988), who Raven Press, New York. demonstrated synergistic effects between the BRADFORD, M.M. 1976. -4 rapid and sensitive method for the quantitation of microgram quantities of cGI-PDE inhibitor cilostazol, a cilostamide protein utilizing the principle of protein-dye derivative, and forskolin. I t is more difficult to binding. ,4nai)lt Biochem 72, 248--254. interpret the findings of Schoeffter et al. (1987), who demonstrated that the cilostamide analogue BRYSOS,S.E. & RODGER,I.W. 1987. Effects of phosphodiesterase inhibitors on normal and chemi.3AJ2 05 enhanced the functional and biochemical cally-skinned isolated airway smooth muscle. Br 3 effects of isoprenaline, while cilostamide, a Pharinacol 92, 673-681. known inhibitor of the cGI-PDE (Beavo 1988), DACSSE, J.-P., CARDOT, a. & MEYER, P. 1983. Specific had no such actions. binding of 3H-guanfacine to z*-adrenoceptors in Rolipram consistendy enhanced the isorat brain. 3 Phurmacol 14, 35-46. prenaline-induced increases in cAMP levels but HAGIWARA, M., ENDO,T. & HIDAKA, H. 1984. Effects of vinpocetine on cyclic nucleotide metabolism in not its relaxant effects in guanfacin-contracted

Combined eferts qf cidenylate rycliise stimulators und PDE inhibitors

PDE inhibitors, rat aorta and platelets

219

1986. Eicosanoid metabolism and P-adrenergic vascular smooth muscle. Biochem Pharmacol 33, mechanisms in coronary arterial smooth muscle : 453-457. B., CADD, potential compartmentation of CAMP. A m 3 Physiol HARRISON, S.A., REIFSNYDER, D.H., GALLIS, 250, C406C412. G.G. & BEAVO, J.A. 1986. Isolation and characteriSCHOEFFTER, P., LUGNIER, C., DEMESY-WAELDELE, F. zation of bovine cardiac muscle cGMP-inhibited phosphodiesterase : a receptor for new cardiotonic & STOCLET, J.C. 1987. Role of cyclic AMP- and drugs. Mol Pharmacol 29, 506-514. cyclic GMP-phosphodiesterases in the control of N.J. 1981. Blood platelet HASLAM, R.J. & CUSACK, cyclic nucleotide levels and smooth muscle tone in receptors for ADP and for adenosine. In: G. rat isolated aorta. Biochem Pharmacol 36, 3965Burnstock (ed.) Purinergic Receptors, pp. 223-285. 3972. SILVER, P.J. 1989. Biochemical aspects of inhibition of Chapman & Hall, London. cardiovascular low (Km) cyclic adenosine monoHIDAKA, H. & ASANO,T. 1976. Human blood platelet phosphate phosphodiesterase. A m 3 Cardiol 63, 3’ :5’-cyclic nucleotide phosphodiesterase. Isolation 2A-8A. of low-Km and high-Km phosphodiesterase. SILVER, P.J., LEPORE,R.E., O’CONNOR,B., LEMP, Biochim Biophys Acta 429, 485-497. B.M., HAMEL,L.T., BENTLEY, R.G. & HARRIS, H . & ENDO,T . 1984. Selective inhibitors of HIDAKA, A.L. 1988. Inhibition of the low Km cyclic AMP three forms of cyclic nucleotide phosphosphodiesterase and activation of the cyclic AMP phodiesterase - basic and potential clinical applisystem in vascular smooth muscle by milrinone. 3 cations. A d v Cyclic Nucleotide Res 16, 245-259. Pharmacol Exp Ther 247, 3442. HIDAKA,H., HAYASHI, H., KOHRI,H., KIMURA,Y., A.W.M., REEVES, M.L. & RINK,T.J. 1988. HOSOKAWA, T., IGAWA,T. & SAITOH,Y. 1979. SIMPSON, Effects of SK&F 94120, an inhibitor of cyclic Selective inhibitor of platelet cyclic adenosine nucleotide phosphodiesterase type 111, on human monophosphate phosphodiesterase, cilostamide, inplatelets. Biochem Pharmacol 37, 2315-2320. hibits platelet aggregation. 3 Pharmacol Exp Ther SOUNESS, J.E., BRAZDIL, R., DIOCEE, B.K. &JORDAN, 211, 2630. HOGESTATT, E.D., ANDERSSON,K.-E. & EDVINSSON, R. 1989. Role of selective cyclic GMP phosphodiesterase inhibition in the myorelaxant actions of L. 1983. Mechanical properties of rat cerebral M & B 22,948, MY-5445, vinpocetine and 1arteries as studied by a sensitive device for recording methyl-3-isobutyl-8-(methylamino)xanthine. Br 3 of mechanical activity in isolated small blood vessels. Pharmacol98, 725-734. Acta Physiol Scand 117, 49-61. T., ISHIKAWA, T., HAGIWARA, M., ONODA, JACKSON, C., BALL,J., PEEL,J., LAWRY, J., GREAVES, TANAKA, H. 1988. Effects of K., ITOH, H. & HIDAKA, F.E. 1989. DN 9693: A phosphoM. & PRESTON, cilostazol, a selective cAMP phosphodiesterase diesterase inhibitor with a platelet membrane effect. inhibitor on the contraction of vascular smooth Thromb Haemost 61, 266269. LUGNIER, C., SCHOEFFTER, P., LEBEC,A., STROUTHOU, muscle. Pharmacology 36, 3 13-320. S.S. & FROJMOVIC, M.M. 1980. Inhibition of E. & STOCLET, J.C. 1986. Selective inhibition of TANG, platelet function by antithrombotic agents which cyclic nucleotide phosphodiesterases of human, selectively inhibit low-Km cyclic 3’,5’-adenosine bovine and rat aorta. Biochem Pharmacol 35, monophosphate phosphodiesterase. 3Lab Clin Med 1743-1751. 95, 241-257. MACPHEE, C.H., HARRISON, S.A. & BEAVO, J.A. 1986. R.V.K. & DIAMOND, J. 1986. Elevation of Immunological identification of the major platelet VEGESNA, cyclic AMP by prostacyclin is accompanied by low-Km cAMP phosphodiesterase : Probable target relaxation of bovine coronary arteries and confor anti-thrombotic agents. Proc Natl Acad Sci traction of rabbit aortic rings. EurJ Pharrnacol128, U S A 83, 666G6663. 25-3 1. RANDRIANTSOA, A,, HEITZ,C. & STOCLET, J.C. 1981. R.E., BURROWS, S.D., KOBYLARZ, D.C., Functional characterization of postjunctional a- WEISHAAR, QUADE, M.M. & EVANS,D.B. 1986. Multiple adrenoceptors in rat aorta. Eur 3 Pharmacol 75, molecular forms of cyclic nucleotide phospho5740. diesterase in cardiac and smooth muscle and in RASCON, A., BELFRAGE, P., LINDGREN, S., ANDERSON, platelets. Biochem Pharmacol 35, 787-800. A., MANGANIELLO, K.-E., STAVENOW, L., NEWMAN, H. & CAWTHORNE, V.C. & DEGERMAN, E. 1990. Purification and YOUNG,P., BERGE,J., CHAPMAN, A. 1989. Novel a,-adrenoceptor antagonists show characterization of the cGMP-inhibited low Km selectivity for a,,-and a,,-adrenoceptor subtypes. phosphodiesterase from bovine (aortic) smooth Eur 3 Pharmacol 168, 381-386. muscle. Submitted. G., GALVAS, P., DISALVO, J. & PAUL,R.J. RUBANYI,