Eur. J Biochem. 207,607-613 (1992)
0FEBS 1992
Cyclic nucleotides and intracellular-calcium homeostasis in human platelets Bernhard BRUNE and Volker ULLRICH Faculty of Biology, University of Konstanz, Federal Republic of Germany (Received February 20/March 30, 1992)
~
EJB 92 0236
The relationship between agonist-sensitive calcium compartments and those discharged by the Ca2+-ATPase inhibitor thapsigargin were studied in human platelets. In this context, calcium mobilization from intracellular pools and manganese influx was investigated in relation to the effect of altered cyclic-nucleotidelevels. For maximal calcium release from intracellular stores, thapsigargin, compared to a receptor agonist like thrombin, requires the platelet’s self-amplification mechanism, known to generate thromboxane A2.With this lipid mediator formed, thapsigargin released calcium and stimulated manganese influx in a manner similar to thrombin. Blocking the thromboxane receptor by addition of sulotroban (BM13.177) or, alternatively, increasing platelet cAMP or cGMP using prostacyclin or sodium nitroprusside, dramatically reduced the ability of thapsigargin to release calcium from intracellular compartments. The same experimental conditions significantly reduced the rate of manganese influx initiated by thapsigargin compared to thrombin. The experiments indicate that thapsigargin-sensitive compartments play only a minor role in inducing manganese influx compared to the receptor-sensitive compartment. Cyclic nucleotides accelerate the redistribution of an agonist-elevated platelet calcium into the thapsigargin-sensitive compartment, from which calcium can be released by inhibition of the Ca2+-ATPase.In human platelets, thapsigargin-induced calcium increase and influx were responsible for only part the calcium release resulting from inhibition of the corresponding ATPase; another part results from the indirect effect of thapsigargin acting via thromboxane-A2-receptor activation. Cyclic nucleotides are therefore an interesting regulatory device which can modify the thapsigargin response by not allowing the self-amplification mechanism of platelets to operate.
Human platelets are stimulated by various agonists which act on specific receptors [l]. Among several activating compounds, one also finds thrombin and thromboxane Az; these initiate an ordered sequence of events associated with an activation of phospholipase C [2] during which the hydrolysis of phosphatidylinositol 4,5-bisphosphate produces at least two second messengers: inositol 1,4,5-trisphosphate, which accounts for a cytosolic-calcium increase, and 1,2-diacylglycerol known to activate protein kinase C [3, 41. These compounds act synergistically, inducing further platelet responses such as platelet aggregation and secretion. An increased level of cytosolic calcium activates phospholipase Az,which concomitantly releases arachidonic acid from membrane phospholipids. Free arachidonate is metabolized via prostaglandin endoperoxides to generate thromboxane AZ.Endoperoxides and thromboxane A2 are potent activators by themselves and serve to amplify the response to weak agonists [5-71. The release of calcium from intracellular stores by the action of inositol 1,4,5-trisphosphate has been demonstrated [3], as well as the central role of calcium during platelet activation [S, 91. Recently, we reported [lo, 111 the existence of two intracellular calcium pools in human platelets, which can be discriminated either by receptor activation or Ca2+-ATPase inhiCorrespondence to B. Brune, University of Konstanz, Faculty of Biology, Universitatsstrasse 8 - 10, Postfach 5560, W-7750 Konstanz, Federal Republic of Germany Fax: +49 7531 882966. Abbreviation. [CaZ+Ii, intracellular calcium concentration
bition using compounds like thapsigargin [12 - 141 or 2,5-di(tbutyl)-l,4-benzohydroquinone[15- 171. Defining calcium release, after receptor stimulation and release of the soluble messenger inositol 1,4,5-trisphosphate, as originating from the classical inositol-l,4,5-trisphosphate-sensitive calcium pool, the calcium traces observed due to thapsigargin should result from prevention of calcium uptake into a thapsigarginsensitive compartment. Blocking of ATPase should result in considerably high passive leakage of calcium during blockage of uptake. We concluded that thapsigargin may release Ca2+ from both inositol-I ,4,5-trisphosphate-sensitiveand inositol1,4,5-trisphosphate-insensitiveintracellular pools. Following the thapsigargin-induced [CaZ‘Ii transient, thrombin was unable to raise [Ca’ +Ii, indicating that thapsigargin mobilizes calcium from the thrombin-responsive store, as long as the self-amplifying system of platelets is intact. With the thromboxane receptor blocked, thapsigargin releases only half the total calcium, with thrombin being able to reIease additional calcium. Interestingly, in the converse experiment, thrombin did not prevent a rise in [Ca2+Iidue to thapsigargin, although applying thrombin a second time had no effect. These observations are consistent with findings recently reported using permeabilized GH4C1 rat pituitary cells [18], permeabilized rat acinar cells [19], NG108-15 and PC12 cells
POI.
In blood platelets and vascular smooth muscle, increases in both cAMP and cGMP by activation of adenylate cyclase or soluble guanylate cyclase cause negative regulation. Increases in cAMP inhibit agonist-induced rises in intracellular
608 calcium and decrease the calcium sensitivity of the activation mechanism [21 -251. Increasing cGMP is also known to inhibit agonist-induced elevation of [Ca2+Ii[26-291. Both cyclic nucleotides mediate the inhibitory effect of the two endothelium-derived regulators, prostacyclin and endothelium-derived relaxing factor. A precise regulation of intracellular calcium homeostasis is vital for platelets since an excess leads to aggregation. Although the major physiological function of platelets is in haemostasis, their importance in thrombosis is clear: they initiate and form part of the occluding thrombus. Therefore, the clinical use of antiplatelet drugs is an attempt to reduce the risk of spontaneous thrombus formation. In the present study, we examined the role of CAMP and cGMP in relation to the two-pool model suggested for human platelets. We mainly focused on the modulation of the thapsigargin response using prostacyclin and sodium nitroprusside by studying the intracellular calcium response as well as the modulation of calcium influx. EXPERIMENTAL PROCEDURES Materials
Thrombin was purchased from Hoffmann-La Roche, Basel, Switzerland. Thapsigargin was delivered by Gibco, Eggenstein, FRG. Radiochemicals were obtained from Du Pont, Dreieich, FRG. Prostacyclin, sodium nitroprusside and BW755C were bought from Sigma Chemie, Deisenhofen, FRG. BMI 3.177 (sulotroban) was from Boehringer Mannheim, Mannheim, FRG. All other materials were as previously described [lo, 111 or were obtained in the highest grade of purity available from local commercial sources. Methods Preparation of platelet-rich plasma and wushed human platelets
Platelet-rich plasma and suspensions of washed human platelets were mainly prepared as outlined previously [ll]. Briefly, trisodium citrate (0.38%, mass/vol.) was added to fresh human blood, and platelet-rich plasma was obtained by centrifugation at 200 x g for 20 min. Platelet-rich plasma was removed, supplemented with prostacyclin (50 nglml) and platelets were separated by centrifugation (1200 x g , 10 min), washed once and resuspended in a Hepes-buffered Tyrode's solution (138 mM NaCl, 0,36 mM NaH,PO,, 2.9 mM KC1, 1 mM MgCl,, 5 mM glucose, 20 mM Hepes, pH 7.4). Incubations (5-8 x lo8 platelets/ml) were performed at 37°C for the times indicated. The endoperoxide/thromboxane receptor was blocked with 100 pM BM13.177 by incubation for 2 min before starting each individual experiment. Prostacyclin and sodium nitroprusside were routinely incubated for 2 min. C'ulciuni mimurements
Calcium measurements in intact platelets were carried out as stated in [lo], resuspending platelets in a nominally calciumfrcc medium. In some experiments, external calcium was removed by adding 0.5 mM EGTA. Briefly, platelets were incubated with 2 pM Fura-2 acetoxymethyl ester at 37°C for 40 min, washed and resuspended in the buffer described above. Changes in fluorescence of Fura-2-loaded platelets at the excitation wavelengths of 335 nm and 362nm, and emission wavelengths above 450 nm were determined following the addition of antagonists and agonists using a Sigma
ZWS-11 dual-wavelength spectrofluorimeter (Biochem, Puchheim, FRG) [30]. The ratios of maximum and minimum fluorescence were determined by the addition of 250 pM digitonin in the presence of 1 mM free calcium or in the presence of 10 mM EGTA (pH 8.5), respectively. Solvents or inhibitors did not affect the method. Prostacyclin, sodium nitroprusside or BM13.177 were incubated for 2 min before adding any agonist. Calcium influx into platelets was measured using the manganese technique described by Sage et al. [31]. This method has previously been employed to show that stimulation of receptor-mediated calcium entry will result in the concomitant stimulation of M n 2 + influx and is based on the observation that Mn2+ binds to Fura-2 and quenches its fluorescence. The use of the two excitation wavelengths, 335 nm and 362 nm, allows one to monitor at calcium-sensitive and calcium-insensitive wavelengths simultaneously. The use of excitation at 362 nm thus allows the selective study of Mn2+ entry without interference to the signal caused by changes in [Ca2+Ii.Control experiments using 5 mM Ni2+ were performed to block manganese entry and to certify influx through so-called receptor-operated channels. Thromboxune-B, formation and rudioimmunoussuy
Formation of thromboxane B2 was measured in connection with the calcium measurements using a radioimmunoassay. The antiserum was obtained from ICN Biomedicals, Inc., Eschwege, FRG, showing a maximal binding of 71%. Samples were processed as previously described [32]. The variation between assays did not exceed 10% in different determinations. Standard curves contained 15 - 2000 pg thromboxane B2. Unstimulated control samples contained 1.5 0.4 ng thromboxane B2/ml platelets (mean & SD, n= 10). After stimulation with various agonists, including thapsigargin, thromboxane B2 values ranged over 70- 140 ng thromboxane B2/ml platelets. Statisticul methods
Results are expressed as mean SD of n determinations of individual experiments from different blood donors. In the case of calcium measurements, only one typical trace of a minimum of five different experiments is shown. Results in all experiments presented were of a magnitude similar to those shown and variability did not exceed 10% of the effect observed. RESULTS Effect of thapsigargin and thrombin on cytosolic calcium
Addition of thapsigargin to Fura-2-loaded human platelets in the presence of extracellular EGTA, with the thromboxane receptor being blocked (addition of BM13.177), caused a relatively slow increase in cytosolic calcium (Fig. 1A). After a minor calcium overshoot, an elevated steady-state plateau level was reached after about 2 min. In this respect, 1 pM thapsigargin was maximally effective. Stimulation with a maximally effective concentration of thrombin at increasing times (2.5 - 60 min) during this plateau phase revealed that still more calcium was released. It is also obvious that there is a progressive decrease in the peak maximum of the thrombin-induced calcium response during this time. However, a significant calcium signal could still be
609
i 0
f1
Thi
TG
1
1
1
1
1
1
1
15
30
L5
60
15
30
L5
60
1
0 5 m M EGTA
-
!
!
J
T
-
w
-
-
25 05mM EGTA
0
25
5
10
TIME, rnin
Fig. 1. Time-dependent intracellular calcium release by thrombin and thapsigargin lntracellular calcium was measured in human platelets resuspended in a nominally calcium free medium, using the fluorescent CaZ indicator Fura-2 as described under Experimental procedures. Extracellular calcium was chelated by adding 0.5 mM EGTA, whereas the thromboxane receptor was blocked by BM13.177 (100 pM). Thapsigargin (TG, 1 pM) and thrombin (Thr, 0.1 Ujml) were employed according to the time schedule indicated at the concentrations given. None of the agents nor their carrier solvents had a direct effect on Fura-2 fluorescence. (A) Treatment of platelets for different time periods with thapsigargin followed by the addition of thrombin; (B) response to thrombin added alone using the same time frame. +
detected after a 1-h exposure to thapsigargin. The additional calcium release seen with thrombin was always transient, with a rapid decay phase that was not significantly affected by increasing the duration of thapsigargin treatment. Fig. 1B shows the response of the same platelets towards thrombin alone in a similar time-course experiment. In the presence of EGTA and BM13.177, the peak maximum of intracellularcalcium release declines significantly with the time of incubation. Compared to the thapsigargin signal, thrombin released calcium more rapidly, without any time delay. At lower time points (2.5 min and 5 rnin), the combined addition of thapsigargin and thrombin released the same initial amount of calcium as did thrombin added alone. At incubation times greater than 30 min, the differences became more obvious, indicating that thrombin releases more calcium than thrombin after incubation with thapsigargin. Cyclic nucleotides modulate thapsigargin-induced calcium increase
The thapsigargin-induced calcium rise is dependent on extracellular calcium as well as on an intact self-amplification mechanism of the platelet suspension (Fig. 2). As in many other cell types, the cytosolic calcium increase, as well as the duration of the calcium signal after cell stimulation, depends on the presence of extracellular calcium. Fig. 2a shows a calcium increase after stimulation with 1 pM thapsigargin and its modulation by prostacyclin and sodium nitroprusside in a nominally calcium-free medium. Thapsigargin caused a sharp rise in cytosolic calcium from a resting level of around 80 nM to about 750 nM [Ca2+Ii,followed by a slow decay in the signal. Addition of either prostacyclin or sodium nitroprusside causes a dramatic reduction of the initial cytosolic calcium increase. The observed calcium signal is much slower to reach an elevated [Ca2+Iiafter several minutes, without any tendency of the signal to decline after-
wards. Fig. 2b reflects that the thapsigargin-induced cytosoliccalcium increase is impaired by the reduction in extracellular calcium. Addition of EGTA causes a reduction in the initial calcium increase after thapsigargin, and the decline of the elevated cytosolic calcium is much faster, reaching an elevated plateau or sustained phase which remains elevated above basal levels. Again, addition of prostacyclin or sodium nitroprusside alter the thapsigargin-induced calcium release dramatically. The rapid initial calcium increase is suppressed and the slow intracellular calcium increase reaches nearly the same elevated plateau phase observed in the absence of prostacyclin or sodium nitroprusside. Fig. 2c describes the response towards thapsigargin in the presence of EGTA in combination with a blocked thromboxane receptor. After the addition of thapsigargin, no rapid calcium-release phase can be observed. The response is slow, reaching an elevated [Ca2+Iiafter several minutes, with no indication of the signal showing a tendency to decline. Addition of prostacyclin or sodium nitroprusside only marginally affect the calcium-release properties of thapsigargin under these conditions. Although the calcium increase is slowed down even further, the end point of the cytosolic calcium increase seems to be similar. Thromboxane-A2 formation is an indication of an active, calcium-dependent phospholipase A2, followed by metabolism of free arachidonate to potent short-lived lipid mediators, a process normally associated with the platelet selfamplification mechanism during cell activation. We were interested in studying the formation of thromboxane BZ,parallel to our calcium measurements, because an increase in calcium and thromboxane formation should be linked. As shown in Table 1, samples were directly withdrawn during Fura-2 measurements and, therefore, refer to the experimental design given in the previous Fig. 2. Time-course experiments revealed that icosanoid formation was already maximal when samples were withdrawn after 4 rnin (data not shown). Evidently, addition of thapsigargin produced significant amounts
610
b 4
'
B
A
1
-A
____
C 0 5mM EGTA lOOuM EM13177
05mM EGTA
I
I E TG[lpMI A PGI211pg/mll B TGllpMI A SNPllWpMl B T G l l p M l
A
h
I\
1
-A
E
1
, B TGllpM1
---- A PGi2llug/mll B TG(1pMI A SNPllWpMl
B TGIlpMI
-A
I B TGllpMI PGI211pg/mll B TGIlpMI A S N P l l W p M I B TGllpM1
_-__ A
Fig. 2. Cyclic nucleotides modulate thapsigargin-induced calcium release. Changes in [Ca' 'Ii were monitored with Fura-2, as indicated under Experimental Procedures. Concentrations of thapsigargin (TG), prostacyclin (PGI') and sodium nitroprusside (SNP) are as shown in the figure, added according to the protocol indicated, in a nominally calcium-free medium (a), after addition of EGTA (b) or with EGTA and BM 13.177 (c).
thromboxane B2; therefore, all other responses were compared to this effect. Blocking the endoperoxide/thromboxane receptor by treatment with BM13.177 dramatically reduced thromboxane levels after employing thapsigargin. Interestingly, as shown in Fig. 2c, the calcium-release properties of thapsigargin are modulated significantly under these conditions. Addition of prostacyclin or sodium nitroprusside for 2 min before stimulation with thapsigargin results in the same dramatic reduction in thromboxane-€3, formation. Again, this situation is somehow reflected (Fig. 2 b) when studying cytosolic-calcium alterations under these conditions. Unstimulated assays contain approximately 1.5 ng thromboxane B,/ ml platelets. In the presence of BM13.177, prostacyclin or sodium nitroprusside, the thromboxane levels ranged only 2 3-fold above reference values, indicating that addition of a thromboxane-receptor blocker, or incubation with prostacyclin or sodium nitroprusside drastically reduced the ability of thapsigargin to cause thromboxane-B, formation. To characterize the effect of CAMP- or cGMP-elevating agents on intracellular calcium homeostasis in relation to the two-pool model, we investigated the redistribution of calcium already released (Fig. 3a and b). Using a platelet suspension in a nominally calcium-free solution with a blocked thromboxane receptor, (prior incubation with BM13.177), addition of thrombin produced a sharp immediate intracellularcalcium increase. After reaching a peak maximum, a slow decline in the Fura-2 signal was observed. Introducing thapsigargin (Fig. 3a, C) transiently releases stored calcium. Increasing the level of cyclic nucleotides (Fig. 3a, B), by adding prostacyclin or sodium nitroprusside after cytosolic calcium had been elevated by thrombin, accelerated the decline in the Fura-2 signal, with sodium nitroprusside being more effective compared to prostacyclin. Interestingly, by applying thapsigargin later, stored calcium was released again to the same extent (peak height) as in control incubations without altered cyclic-nucleotide levels. The same experimental setup described in Fig. 3a was used in Fig. 3b, with the exception of carrying out the experiments in the presence of a blocked thromboxane receptor in combination with all extracellular calcium being removed by addition of EGTA. After stimulation with thrombin, we observed a rapid intracellular-calcium increase followed by a relatively fast and nearly complete calcium uptake into intra-
cellular stores. Stored calcium can be released by subsequent addition of thapsigargin. In a situation excluding extracellular-calcium influx, the intracellular-calcium increase after thrombin shows the same peak height, but resequestration is much faster. Addition of prostacyclin or sodium nitroprusside (Fig. 3b, B) further accelerates calcium reuptake, again with the tendency of sodium nitroprusside to be more potent compared to prostacyclin. Applying thapsigargin (Fig. 3b, C) again released stored calcium, without any differences between individual assays. For comparison, we included the calcium trace produced by thapsigargin in the presence of BMl3.177 and EGTA when added alone, without a prior incubation employing thrombin as the first stimulus. Thapsigargin alone only releases about half of the calcium compared to thrombin. Cyclic nucleotides modulate thapsigargin-induced calcium influx
Calcium influx into platelets was measured using the manganese technique described by Sage and coworkers [31]. Stimulation of a receptor-operated calcium channel will result in stimulation of Mn2 influx associated with quenching of the dye Fura-2. Exposure of platelets to 1 pM thapsigargin augmented the rate of M n 2 + entry when compared to the rate measured during the control period before thapsigargin addition (Fig. 4a, trace 1). Blocking the thromboxane receptor with the receptor antagonist BM13.177 significantly reduces Mn2+ entry after adding thapsigargin (Fig. 4a, trace 2 versus 1). Inclusion of prostacyclin with BM13.177 abolishes the residual manganese entry (Fig. 4a, trace 3) after addition of thapsigargin. The same inhibitory activity of prostacyclin was demonstrated when BM13.177 was omitted (data not shown). In the same set of experiments we investigated the effect of sodium nitroprusside on thapsigargin-induced Mn2 influx, as shown in Fig. 4b. As indicated, thapsigargin-initiated M n 2 + influx is dramatically reduced in the presence of BM13.177 and is totally inhibited by further addition of sodium nitroprusside. When BM13.177 was omitted from the experiment, the same inhibitory activity of sodium nitroprusside was observed (data not shown). In this respect, there is no difference between either CAMP or cGMP. For comparison, we studied the thrombin-mediated Mn 2 + influx as presented in Fig. 5. Fig. 5a shsws that addition of thrombin and +
+
61 1
a
b 0 5mM EGTA lOOuM BM13177
100pM L
W U
z
W
10 U
3
U
2min
-A
ThrlOlUlmil
.___ A ThiIOlUlnll
A ThrlOlUlmll
E
-A
ThrIOlU/mil A Thr[OlU/mll ......__A:ThrlOlU/mII A TG[lpMI
I C TG[lpMI
E PG1211pg/mil E SNP1100pMl
____
C TG[lpMI C TGIlpMI
B. I C TGIlpMI B-PGl,Ilpg/mil C TGIlpM1 B .SNP[lOOpMI C TGIlpMI
Fig. 3. Cyclic nucleotides accelerate calcium redistribution into thapsigargin-sensitive calcium compartments. Fura-2-loaded platelets were challenged with various compounds at indicated concentrations according to the protocol given, in the presence of either BM13.177 (a) or BM13.177 plus EGTA (b). Intracellular calcium was measured as described under Experimental Procedures. TG, thapsigargin; PG12, prostacyclin; SNP, sodium nitroprusside; Thr, thrombin.
a
thrombin compared to thapsigargin, an inhibitor of the Ca2 ATPase, especially in a situation not allowing the self-amplification mechanism to be intact. Calculating the initial rate of Fura-2 quenching after using thapsigargin as an agonist (100% control value), addition of BM13.177 reduces this initial rate to values below 20% (80% inhibition). Investigating calcium entry and calculating the rate of [Ca2+Iiincrease during the first 1 0 s after Ca2+ addition revealed that the initial rate after addition of thapsigargin (100% control value) was only reduced by about 40% in the presence of BM13.111. These results indicate differences between [Ca2'Ii rise and Mn2+ entry. + -
0 0 DISCUSSION @ @ @)
I B TGIlpMI A BM13177IlOOpMI 8 - TGIlpMI .. ..-.- A BM13177[XX)pMI 8 T G I l p M I PG12 [lpg/rnll
-A
@ @ @)
I 6 TGflpMI A BM131771WOpMI 8 TGlIpMI -.-_...-. A EM13177[XX)pM1 B T G I l p M I SNP I100pMI
-A.
Fig. 4. BM13.177, prostacyclin and sodium nitroprusside inhibit Fura2 fluorescence in the presence of Mn2+. The fluorescence emission (excitation wavelength, 362 nm) of Fura-2-loaded platelets was recorded in the presence of 500 pM extracellular M n Z + ,as indicated under Experimental Procedures. At the various points indicated, different additions at the indicated concentrations were made. TG, thapsigargin; PG12, prostacyclin; SNP, sodium nitroprusside.
thapsigargin resulted in a similar slope of manganese influx. Influx was maximal and could not be further increased by subsequent addition of the other corresponding agonist, thrombin or thapsigargin. As indicated in Fig. 5b, addition of the thromboxane-receptor antagonist BMI 3.177 significantly inhibits thapsigargin-induced manganese influx, whereas thrombin-initiated manganese influx is not affected (Fig. 5b, compare traces 1 and 2). The slow Mn2+ influx caused by thapsigargin in the presence of BM13.177 is dramatically enhanced by the addition of a receptor agonist like thrombin (Fig. 5b, C). Therefore, manganese-influx experiments also reveal differences in the action of a receptor agonist like
Calcium release from intracellular stores and calcium influx Previous findings revealed the structural separation of a receptor-sensitive calcium-release compartment versus a compartment showing a sensitivity towards a Ca2+-ATPaseinhibitor like thapsigargin [I I]. Although earlier data suggested that thapsigargin empties the same calcium pool released by inositol 1,4,5-trisphosphate 1321, it became evident that thapsigargin also affects inositol-l,4,5-trisphosphate-insensitive compartments [ l l , 18-20, 33, 341. In platelets with an intact self-amplification system, thapsigargin. by acting not only as a Ca2+-ATPaseinhibitor but also indirectly as a receptor agonist, produces calcium traces similar to thrombin. This suggests that a receptor agonist and a Ca2+-ATPase inhibitor are indistinguishable in releasing calcium from intracellular stores. With a blocked thromboxane receptor, the agonistinduced calcium increases were always larger than those evoked by thapsigargin (Fig. 1). However, the magnitude of thapsigargin-induced Ca2+ increases varies considerably among different cell types, and in platelets it is also influenced by the endogenous formation of thromboxane A2. The time course of the progressive decay of the maximal thrombininduced Ca2 transient after thapsigargin treatment indicates that the agonist-sensitive compartment also shows a leakage +
61 2
b
a
@ @
B TG[lpMI I B Thr[OlU/mll I
-A
C Thr[OlU/rnll C TGIlpMI
B TGIlpMI @ -A:BM13177[1WpMi @ ----A BM13177[100pMI B ThrlOlU/rnll
C-ThrIOlUlrnll C TGllpMI
Fig. 5. Thapsigargin- and thrombin-induced manganese influx and its modulation by BM13.177. Mn2+ influx in platelets was measured as described under Experimental Procedures and explained in Fig. 4. At the indicated points, either BMI 3.1 77, thapsigargin (TG) or thrombin (Thr) were added at the concentrations indicated.
which is considerably slower than the leakage of the Table 1. Thrornboxane-BZ formation in washed human platelets. thapsigdrgin-sensitive compartment. This effect can not be Washed human platelets (5 -8 x lo8 platelets/ml) were taken during explained simply by removal of Ca2+ from the cells because original Furd-2 recordings. Samples were processed according to ExProcedures and thromboxane BZwas determined accordthe thrombin signal is more stable. It might indicate that the perimental ingly. Results are expressed as percentage of a control value in each thapsigdrgin-sensitive compartment may serve as a reservoir experiment, set as loo%, expressed as percentage of thromboxane-B2 for refilling the agonist-sensitive part. In this respect, the re- (TxB,) formation compared to the response towards thapsigargin. sults fit a proposal of Menniti and colleagues [35] describing All samples used 1 pM thapsigargin. Means 5 SD from four experthe inositol-I ,4,5-trisphosphate-sensitivestore as one com- iments are given. Samples were taken according to the experimental partment involved in Ca2' uptake and another compartment design presented in the corresponding figure, always 4 min after the mainly involved with Ca2 release. Our results demonstrate addition of thapsigargin. Prostacyclin (PG12) sodium nitroprusside that despite a prolonged incubation with thapsigargin in the (SNP) and BM13.177 were incubated for 2 min. presence of extracellular EGTA, the structural separation of Sample TxB~ two calcium-release compartments remains intact. formation Studying the functional relationship between Ca2' entry and intracellular-Ca2+pools in human platelets, thapsigargin % and thrombin produce equivalent Mn2+influx (Fig. 5 ) as long 100 as the self-amplification system remains intact. Preventing the Control 100 pM BM13.177 1.3 k 0.52 formation of thromboxane A2 reduces thapsigargin-induced 1 pg/ml prostacyclin 0.6 k 0.47 M n2+influx compared to that stimulated by thrombin. Intra- 100 pM sodium nitroprusside 2.1 0.50 cellular-calcium mobilization [I I], as well as the Ca2+-influx studies, indicate that thapsigdrgin alone, without endogenous formation of thromboxane A,, releases the part of the intracellular calcium associated with only a marginal M n 2 +influx, Cyclic nucleotides modulate thapsigargin-inducedcalcium although showing only partial inhibition of C a 2 +entry (Ca2' release overshoot). Our results revealed that thapsigargin-sensitive As previously reported and also shown here, thapsigargin intracellular-calcium compartments only play a minor role in opening channels allowing manganese influx (Fig. 5), com- effectively releases calcium and causes thromboxane-B, forpared to a classical receptor agonist like thrombin. Therefore, mation (Fig. 2; Table 1) by the ability of platelets to amplify in platelets, not only capacitative Ca2+ entry [37] originating the signal, resulting in the formation of a feed-forward actifrom empty thapsigargin-sensitive compartments, but also an vation signal. Cutting off the self-amplification system by additional component mediated by receptor activation and blocking the thromboxane receptor inhibits thapsigargin-inemptying of the related store seems to be involved in allowing duced thromboxane-B2 formation and modifies calcium reMn2+ influx. Similar observations, that simply emptying the lease significantly. Applying agents like prostacyclin or soinositol-l,4,5-trisphosphate-sensitive Ca2 stores did not lead dium nitroprusside, known to generate cAMP or cGMP inside to a significant stimulation of Ca2 entry, have already been the cell by activation of adenylate cyclase or soluble guanylate observed using hepatocytes [37, 381. Taking into account the cyclase, respectively, caused a similar modification of Ca2+-overshootmodel compared to M n 2 + entry, one should thapsigargin-induced calcium signals due to BM13.177. consider the possibility that emptying thapsigargin-sensitive Measuring thromboxane-Bz formation indicated that cAMP stores may open a specific Ca2' channel, not permeable to and cGMP inhibited the endogenous formation of this lipid Mn2' [39]. This would suggest the existence of two Ca2'- mediator and thereby blocked the self-amplification mechanentry pathways, with only one allowing the influx of M n 2 + . ism. Inhibition of phospholipase-C-mediated cell activation +
+
+
61 3 has been described for both cAMP and cGMP [40,41]. Therefore, prostacyclin and sodium nitroprusside mainly modulate the action of thapsigargin originating from the secondary ability of the Ca2+-ATPaseinhibitor when thapsigargin also acts as an indirect receptor agonist. Small differences between BM13.177 and prostacyclin/sodium-nitroprussidemay originate from calcium extrusion or indicate incomplete inhibition of the thromboxane receptor. Interestingly, thapsigargin still increased the intracellular calcium to an elevated plateau in the presence of the cyclic-nucleotide-forming compounds, indicating that this new steady state is not antagonized by plasma-membrane pump activation. Considering the twopool-compartment model, further experiments (Fig. 3) revealed that cyclic nucleotides stimulate calcium reuptake into the thapsigargin-sensitive intracellular store. This effect is more pronounced in the presence of extracellular calcium, with sodium nitroprusside being more effective than prostacyclin. Additional results (Fig. 4) support the concept that cAMP and especially cGMP are potent inhibitors of C a 2 + influx, as studied using human platelets [28, 421. This is also evident using thapsigargin to induce manganese influx as an indication of Ca2+ entry. The same results were observed employing thrombin (data not shown). Furthermore, the accelerated reversal of an elevated cytosolic-calcium level reported by MacIntyre et al. [27] was confirmed, indicating that uptake into the thapsigargin-sensitive compartment occurs, obviously not associated with calcium extrusion. Therefore, one can speculate that the thapsigargin-sensitive Ca2 ATPase is regulated by cAMP/cGMP taking part in the complex, an efficient way of cyclic nucleotides inhibiting complete platelet activation [43]. Our experiments reveal the interference of the cyclic nucleotides CAMP and cGMP with various calcium signals evoked by thapsigargin. Increased levels of cyclic nucleotides inhibit the self-amplification system of human platelets, not allowing the formation of thromboxane B2 after addition of thapsigargin. The Iack of a positive-feedback loop alters the calcium-release properties of thapsigargin compared to thrombin. By blocking the thromboxane receptor or adding prostacyclin or sodium nitroprusside, thapsigargin releases only part of the intracellular calcium compared to a normal receptor agonist like thrombin. Manganese influx, used as a means to monitor C a 2 + entry, is inhibited, and cyclic nucleotides accelerate the redistribution of cytosolic calcium into the thapsigargin-sensitive calcium compartment.
9. 10. 11 12 13 14 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.
+ -
The study was supported by the Deutsche Forschungsgemeinschaft (SFB 156/A4). The expert technical assistance of B. Diewald is gratefully acknowledged.
REFERENCES 1. Lapetina, E. G. (1986) in Phosphoinositides and receptor mechanisms (Putney, J. W., Jr, ed.) pp. 271 -286, Alan R. Liss, Inc., NY. 2. Lapetina, E. G. (1 983) L(fe Sci. 32, 2069 - 20823. 3. Berridge, M. J. & Irvine, R. F. (1984) Nature312, 315-321. 4. Nishizuka, Y. (1984) Nature 308, 693 - 698. 5. Flower, R. J. & Blackwell, G. J. (1979) Biochem. Pharmacol. 25, 285 - 291. 6. Hamberg, M., Svensson, J. & Samuelsson, B. (1975) Proc. Natl Acad. Sci. 72, 2994-2998. 7. Siess, W., Cuatrecasas, P. & Lapetina, E. G. (1983) J . Bid. Chem. 258,4683 -4686. 8. Rink, T. J. & Hallam, T. J. (1984) Trends Biochem. Sci. 376,215219.
25. 26. 27. 28 29. 30.
Rink, T. J. (1986) Agents Actions Suppl. 20, 141-169. Brune, B. & Ullrich, V. (1991) FEBS Lett. 284. 1-4. Brune, B. & Ullrich, V. (1991) J . Bid. Chem. 266,19 232-19 237 Thastrup, O., Dawson, A. P., Scharff, O., Foder, B., Cullen, P. J., Drobak, B. K., Bjerrrum, P. J., Christensen, S. B. & Hanley, M. R.(1989) Agents Actions 27, 17-23. Thastrupp, O., Foder, B. & Scharff, 0. (1987) Biochem. Biophys. Res. Commun. 142, 654- 660. Jackson, T. R., Patterson, S. L. & Thastrup, 0. (1988) Biochem. J . 253, 81 - 86. Moore, G. A., McConkey, D. J., Kass, G . E. N., O'Brien, P. J. & Orrenius, S. (1987) FEBS Leu. 224, 331 -336. Kass, G. E. N., Duddy, S., Moore, G . A. & Orrenius, S. (1989) J . Bid. Chem. 264, 15 192- 15 198. Oldershaw, K. A. &Taylor, C. W. (1990) FEBS Lett. 274, 21421 6. Koshiyma, H. & Tashjian, A. H., Jr (1991) Biochem. Biophys. Res. Commun. 177, 553 - 558. Menniti, F. S., Bird, G. St. J., Takemura, H., Thastrup, O., Potter, B. V. L. & Putney, J. W., Jr (1991) J . Biol. Chem. 266,13 64613 653. Takemura, H., Ohshika, H., Yokosawa, N., Oguma, K. & Thastrup, 0. (1991) Biochem. Biophys. Res. Commun. 180, 1518- 1526. Yamanishi, J., Kawahara, Y. & Fukuzuki, H. (1983) Thromh. Res. 32, 183-188. Knight, D. E. & Scrutton, M. C. (1984) Nature 309, 66-68. Bushfield, M., McNicoI, A. & MacIntyre, D. E. (1985) Biochem. J . 232,261 - 27 1. Pannochia, A. & Hardisty, R. M. (1985) Biochem. Biophys. Res. Commun. 127,339-345. Sage, S. 0. &Rink, T. J. (1985) FEBS Lett. 188, 135-140. Kawahara, Y., Yamanishi, J. & Fukuzaki, H. (1984) Thromh. Res. 33, 203 - 209. MacIntyre, D. E., Bushfield, M. & Shaw, A. M. (1985) FEBS Lett. 188, 383 - 388. Morgan, R. 0. & Newby, A. C. (1989) Biochem. J . 258, 447454. Brune, B. & Ullrich, V. (1991) Mol. Pharmacol. 39,671 -678. Grynkiewicz, G., Poenie, M. & Tsein, R. Y. (1985) J. Biol. Chem. 260, 3440 3450. Sage, S. O., Merritt, J. E., Hallam, T. J. & Rink, T. J. (1989) Biochem. J . 257,923 - 926. Jackson, T. R., Patterson, S. I., Thastrupp, 0. & Hanly, M. R. (1988) Biochem. J . 253,81-86. Thastrup, O., Cullen, P. J., Drobak, B. K., Hanley, M. R. & Dawson, A. P. (1990) Proc. Natl Acad. Sci. USA 87, 24662470. Ely. J. A., Ambroz, C., Baukal, A. J., Brogger Christensen, S., Balla, T. & Catt, K. J. (1991) J . Biol. Chem. 266, 1863518641. Menniti, F. S., Bird, G. St. J., Takemura, H., Thastrup, O., Potter, B. V. L. & Putney, J. W., Jr (1991) J . Biol. Chem. 266,13 64613 653. Putney, J. W. Jr (1990) Cell Calcium 11, 611 -624. Kass, G. E. N., Llopis, J., Chow, S. C., Duddy, S. K. & Orrenius, S. (1990) J . Bid. Chem. 265, 17 486- 17 492. Hansen, C. A., Yang, L. &Williamson, J. R. (1991) J . B i d . Chern. 266, 18 573 - 18 579. Neher, E. (1 992) Nature 335, 298 - 299. Billah, M. M., Lapetina, E. G. & Cuatrecasas, P. (1979) Biochem. Biophys. Res. Commun. 90,92 - 98. Nakashima, S., Tohmatsu, T., Hattori, H., Okano, Y. & Nozawa, Y. (1986) Biochem. Biophys. Res. Commun. 135,1099-1104. Simon, M. F. & Chap, H. (1989) J . Cardiovasc. Pharmacol. 14, S106 S110. Rink, T. J. & Sage, S. 0. (1990) Annu. Rev. Physiol. 52, 431 490. ~
31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.
~
43.
0FEBS 1992
Cyclic nucleotides and intracellular-calcium homeostasis in human platelets Bernhard BRUNE and Volker ULLRICH Faculty of Biology, University of Konstanz, Federal Republic of Germany (Received February 20/March 30, 1992)
~
EJB 92 0236
The relationship between agonist-sensitive calcium compartments and those discharged by the Ca2+-ATPase inhibitor thapsigargin were studied in human platelets. In this context, calcium mobilization from intracellular pools and manganese influx was investigated in relation to the effect of altered cyclic-nucleotidelevels. For maximal calcium release from intracellular stores, thapsigargin, compared to a receptor agonist like thrombin, requires the platelet’s self-amplification mechanism, known to generate thromboxane A2.With this lipid mediator formed, thapsigargin released calcium and stimulated manganese influx in a manner similar to thrombin. Blocking the thromboxane receptor by addition of sulotroban (BM13.177) or, alternatively, increasing platelet cAMP or cGMP using prostacyclin or sodium nitroprusside, dramatically reduced the ability of thapsigargin to release calcium from intracellular compartments. The same experimental conditions significantly reduced the rate of manganese influx initiated by thapsigargin compared to thrombin. The experiments indicate that thapsigargin-sensitive compartments play only a minor role in inducing manganese influx compared to the receptor-sensitive compartment. Cyclic nucleotides accelerate the redistribution of an agonist-elevated platelet calcium into the thapsigargin-sensitive compartment, from which calcium can be released by inhibition of the Ca2+-ATPase.In human platelets, thapsigargin-induced calcium increase and influx were responsible for only part the calcium release resulting from inhibition of the corresponding ATPase; another part results from the indirect effect of thapsigargin acting via thromboxane-A2-receptor activation. Cyclic nucleotides are therefore an interesting regulatory device which can modify the thapsigargin response by not allowing the self-amplification mechanism of platelets to operate.
Human platelets are stimulated by various agonists which act on specific receptors [l]. Among several activating compounds, one also finds thrombin and thromboxane Az; these initiate an ordered sequence of events associated with an activation of phospholipase C [2] during which the hydrolysis of phosphatidylinositol 4,5-bisphosphate produces at least two second messengers: inositol 1,4,5-trisphosphate, which accounts for a cytosolic-calcium increase, and 1,2-diacylglycerol known to activate protein kinase C [3, 41. These compounds act synergistically, inducing further platelet responses such as platelet aggregation and secretion. An increased level of cytosolic calcium activates phospholipase Az,which concomitantly releases arachidonic acid from membrane phospholipids. Free arachidonate is metabolized via prostaglandin endoperoxides to generate thromboxane AZ.Endoperoxides and thromboxane A2 are potent activators by themselves and serve to amplify the response to weak agonists [5-71. The release of calcium from intracellular stores by the action of inositol 1,4,5-trisphosphate has been demonstrated [3], as well as the central role of calcium during platelet activation [S, 91. Recently, we reported [lo, 111 the existence of two intracellular calcium pools in human platelets, which can be discriminated either by receptor activation or Ca2+-ATPase inhiCorrespondence to B. Brune, University of Konstanz, Faculty of Biology, Universitatsstrasse 8 - 10, Postfach 5560, W-7750 Konstanz, Federal Republic of Germany Fax: +49 7531 882966. Abbreviation. [CaZ+Ii, intracellular calcium concentration
bition using compounds like thapsigargin [12 - 141 or 2,5-di(tbutyl)-l,4-benzohydroquinone[15- 171. Defining calcium release, after receptor stimulation and release of the soluble messenger inositol 1,4,5-trisphosphate, as originating from the classical inositol-l,4,5-trisphosphate-sensitive calcium pool, the calcium traces observed due to thapsigargin should result from prevention of calcium uptake into a thapsigarginsensitive compartment. Blocking of ATPase should result in considerably high passive leakage of calcium during blockage of uptake. We concluded that thapsigargin may release Ca2+ from both inositol-I ,4,5-trisphosphate-sensitiveand inositol1,4,5-trisphosphate-insensitiveintracellular pools. Following the thapsigargin-induced [CaZ‘Ii transient, thrombin was unable to raise [Ca’ +Ii, indicating that thapsigargin mobilizes calcium from the thrombin-responsive store, as long as the self-amplifying system of platelets is intact. With the thromboxane receptor blocked, thapsigargin releases only half the total calcium, with thrombin being able to reIease additional calcium. Interestingly, in the converse experiment, thrombin did not prevent a rise in [Ca2+Iidue to thapsigargin, although applying thrombin a second time had no effect. These observations are consistent with findings recently reported using permeabilized GH4C1 rat pituitary cells [18], permeabilized rat acinar cells [19], NG108-15 and PC12 cells
POI.
In blood platelets and vascular smooth muscle, increases in both cAMP and cGMP by activation of adenylate cyclase or soluble guanylate cyclase cause negative regulation. Increases in cAMP inhibit agonist-induced rises in intracellular
608 calcium and decrease the calcium sensitivity of the activation mechanism [21 -251. Increasing cGMP is also known to inhibit agonist-induced elevation of [Ca2+Ii[26-291. Both cyclic nucleotides mediate the inhibitory effect of the two endothelium-derived regulators, prostacyclin and endothelium-derived relaxing factor. A precise regulation of intracellular calcium homeostasis is vital for platelets since an excess leads to aggregation. Although the major physiological function of platelets is in haemostasis, their importance in thrombosis is clear: they initiate and form part of the occluding thrombus. Therefore, the clinical use of antiplatelet drugs is an attempt to reduce the risk of spontaneous thrombus formation. In the present study, we examined the role of CAMP and cGMP in relation to the two-pool model suggested for human platelets. We mainly focused on the modulation of the thapsigargin response using prostacyclin and sodium nitroprusside by studying the intracellular calcium response as well as the modulation of calcium influx. EXPERIMENTAL PROCEDURES Materials
Thrombin was purchased from Hoffmann-La Roche, Basel, Switzerland. Thapsigargin was delivered by Gibco, Eggenstein, FRG. Radiochemicals were obtained from Du Pont, Dreieich, FRG. Prostacyclin, sodium nitroprusside and BW755C were bought from Sigma Chemie, Deisenhofen, FRG. BMI 3.177 (sulotroban) was from Boehringer Mannheim, Mannheim, FRG. All other materials were as previously described [lo, 111 or were obtained in the highest grade of purity available from local commercial sources. Methods Preparation of platelet-rich plasma and wushed human platelets
Platelet-rich plasma and suspensions of washed human platelets were mainly prepared as outlined previously [ll]. Briefly, trisodium citrate (0.38%, mass/vol.) was added to fresh human blood, and platelet-rich plasma was obtained by centrifugation at 200 x g for 20 min. Platelet-rich plasma was removed, supplemented with prostacyclin (50 nglml) and platelets were separated by centrifugation (1200 x g , 10 min), washed once and resuspended in a Hepes-buffered Tyrode's solution (138 mM NaCl, 0,36 mM NaH,PO,, 2.9 mM KC1, 1 mM MgCl,, 5 mM glucose, 20 mM Hepes, pH 7.4). Incubations (5-8 x lo8 platelets/ml) were performed at 37°C for the times indicated. The endoperoxide/thromboxane receptor was blocked with 100 pM BM13.177 by incubation for 2 min before starting each individual experiment. Prostacyclin and sodium nitroprusside were routinely incubated for 2 min. C'ulciuni mimurements
Calcium measurements in intact platelets were carried out as stated in [lo], resuspending platelets in a nominally calciumfrcc medium. In some experiments, external calcium was removed by adding 0.5 mM EGTA. Briefly, platelets were incubated with 2 pM Fura-2 acetoxymethyl ester at 37°C for 40 min, washed and resuspended in the buffer described above. Changes in fluorescence of Fura-2-loaded platelets at the excitation wavelengths of 335 nm and 362nm, and emission wavelengths above 450 nm were determined following the addition of antagonists and agonists using a Sigma
ZWS-11 dual-wavelength spectrofluorimeter (Biochem, Puchheim, FRG) [30]. The ratios of maximum and minimum fluorescence were determined by the addition of 250 pM digitonin in the presence of 1 mM free calcium or in the presence of 10 mM EGTA (pH 8.5), respectively. Solvents or inhibitors did not affect the method. Prostacyclin, sodium nitroprusside or BM13.177 were incubated for 2 min before adding any agonist. Calcium influx into platelets was measured using the manganese technique described by Sage et al. [31]. This method has previously been employed to show that stimulation of receptor-mediated calcium entry will result in the concomitant stimulation of M n 2 + influx and is based on the observation that Mn2+ binds to Fura-2 and quenches its fluorescence. The use of the two excitation wavelengths, 335 nm and 362 nm, allows one to monitor at calcium-sensitive and calcium-insensitive wavelengths simultaneously. The use of excitation at 362 nm thus allows the selective study of Mn2+ entry without interference to the signal caused by changes in [Ca2+Ii.Control experiments using 5 mM Ni2+ were performed to block manganese entry and to certify influx through so-called receptor-operated channels. Thromboxune-B, formation and rudioimmunoussuy
Formation of thromboxane B2 was measured in connection with the calcium measurements using a radioimmunoassay. The antiserum was obtained from ICN Biomedicals, Inc., Eschwege, FRG, showing a maximal binding of 71%. Samples were processed as previously described [32]. The variation between assays did not exceed 10% in different determinations. Standard curves contained 15 - 2000 pg thromboxane B2. Unstimulated control samples contained 1.5 0.4 ng thromboxane B2/ml platelets (mean & SD, n= 10). After stimulation with various agonists, including thapsigargin, thromboxane B2 values ranged over 70- 140 ng thromboxane B2/ml platelets. Statisticul methods
Results are expressed as mean SD of n determinations of individual experiments from different blood donors. In the case of calcium measurements, only one typical trace of a minimum of five different experiments is shown. Results in all experiments presented were of a magnitude similar to those shown and variability did not exceed 10% of the effect observed. RESULTS Effect of thapsigargin and thrombin on cytosolic calcium
Addition of thapsigargin to Fura-2-loaded human platelets in the presence of extracellular EGTA, with the thromboxane receptor being blocked (addition of BM13.177), caused a relatively slow increase in cytosolic calcium (Fig. 1A). After a minor calcium overshoot, an elevated steady-state plateau level was reached after about 2 min. In this respect, 1 pM thapsigargin was maximally effective. Stimulation with a maximally effective concentration of thrombin at increasing times (2.5 - 60 min) during this plateau phase revealed that still more calcium was released. It is also obvious that there is a progressive decrease in the peak maximum of the thrombin-induced calcium response during this time. However, a significant calcium signal could still be
609
i 0
f1
Thi
TG
1
1
1
1
1
1
1
15
30
L5
60
15
30
L5
60
1
0 5 m M EGTA
-
!
!
J
T
-
w
-
-
25 05mM EGTA
0
25
5
10
TIME, rnin
Fig. 1. Time-dependent intracellular calcium release by thrombin and thapsigargin lntracellular calcium was measured in human platelets resuspended in a nominally calcium free medium, using the fluorescent CaZ indicator Fura-2 as described under Experimental procedures. Extracellular calcium was chelated by adding 0.5 mM EGTA, whereas the thromboxane receptor was blocked by BM13.177 (100 pM). Thapsigargin (TG, 1 pM) and thrombin (Thr, 0.1 Ujml) were employed according to the time schedule indicated at the concentrations given. None of the agents nor their carrier solvents had a direct effect on Fura-2 fluorescence. (A) Treatment of platelets for different time periods with thapsigargin followed by the addition of thrombin; (B) response to thrombin added alone using the same time frame. +
detected after a 1-h exposure to thapsigargin. The additional calcium release seen with thrombin was always transient, with a rapid decay phase that was not significantly affected by increasing the duration of thapsigargin treatment. Fig. 1B shows the response of the same platelets towards thrombin alone in a similar time-course experiment. In the presence of EGTA and BM13.177, the peak maximum of intracellularcalcium release declines significantly with the time of incubation. Compared to the thapsigargin signal, thrombin released calcium more rapidly, without any time delay. At lower time points (2.5 min and 5 rnin), the combined addition of thapsigargin and thrombin released the same initial amount of calcium as did thrombin added alone. At incubation times greater than 30 min, the differences became more obvious, indicating that thrombin releases more calcium than thrombin after incubation with thapsigargin. Cyclic nucleotides modulate thapsigargin-induced calcium increase
The thapsigargin-induced calcium rise is dependent on extracellular calcium as well as on an intact self-amplification mechanism of the platelet suspension (Fig. 2). As in many other cell types, the cytosolic calcium increase, as well as the duration of the calcium signal after cell stimulation, depends on the presence of extracellular calcium. Fig. 2a shows a calcium increase after stimulation with 1 pM thapsigargin and its modulation by prostacyclin and sodium nitroprusside in a nominally calcium-free medium. Thapsigargin caused a sharp rise in cytosolic calcium from a resting level of around 80 nM to about 750 nM [Ca2+Ii,followed by a slow decay in the signal. Addition of either prostacyclin or sodium nitroprusside causes a dramatic reduction of the initial cytosolic calcium increase. The observed calcium signal is much slower to reach an elevated [Ca2+Iiafter several minutes, without any tendency of the signal to decline after-
wards. Fig. 2b reflects that the thapsigargin-induced cytosoliccalcium increase is impaired by the reduction in extracellular calcium. Addition of EGTA causes a reduction in the initial calcium increase after thapsigargin, and the decline of the elevated cytosolic calcium is much faster, reaching an elevated plateau or sustained phase which remains elevated above basal levels. Again, addition of prostacyclin or sodium nitroprusside alter the thapsigargin-induced calcium release dramatically. The rapid initial calcium increase is suppressed and the slow intracellular calcium increase reaches nearly the same elevated plateau phase observed in the absence of prostacyclin or sodium nitroprusside. Fig. 2c describes the response towards thapsigargin in the presence of EGTA in combination with a blocked thromboxane receptor. After the addition of thapsigargin, no rapid calcium-release phase can be observed. The response is slow, reaching an elevated [Ca2+Iiafter several minutes, with no indication of the signal showing a tendency to decline. Addition of prostacyclin or sodium nitroprusside only marginally affect the calcium-release properties of thapsigargin under these conditions. Although the calcium increase is slowed down even further, the end point of the cytosolic calcium increase seems to be similar. Thromboxane-A2 formation is an indication of an active, calcium-dependent phospholipase A2, followed by metabolism of free arachidonate to potent short-lived lipid mediators, a process normally associated with the platelet selfamplification mechanism during cell activation. We were interested in studying the formation of thromboxane BZ,parallel to our calcium measurements, because an increase in calcium and thromboxane formation should be linked. As shown in Table 1, samples were directly withdrawn during Fura-2 measurements and, therefore, refer to the experimental design given in the previous Fig. 2. Time-course experiments revealed that icosanoid formation was already maximal when samples were withdrawn after 4 rnin (data not shown). Evidently, addition of thapsigargin produced significant amounts
610
b 4
'
B
A
1
-A
____
C 0 5mM EGTA lOOuM EM13177
05mM EGTA
I
I E TG[lpMI A PGI211pg/mll B TGllpMI A SNPllWpMl B T G l l p M l
A
h
I\
1
-A
E
1
, B TGllpM1
---- A PGi2llug/mll B TG(1pMI A SNPllWpMl
B TGIlpMI
-A
I B TGllpMI PGI211pg/mll B TGIlpMI A S N P l l W p M I B TGllpM1
_-__ A
Fig. 2. Cyclic nucleotides modulate thapsigargin-induced calcium release. Changes in [Ca' 'Ii were monitored with Fura-2, as indicated under Experimental Procedures. Concentrations of thapsigargin (TG), prostacyclin (PGI') and sodium nitroprusside (SNP) are as shown in the figure, added according to the protocol indicated, in a nominally calcium-free medium (a), after addition of EGTA (b) or with EGTA and BM 13.177 (c).
thromboxane B2; therefore, all other responses were compared to this effect. Blocking the endoperoxide/thromboxane receptor by treatment with BM13.177 dramatically reduced thromboxane levels after employing thapsigargin. Interestingly, as shown in Fig. 2c, the calcium-release properties of thapsigargin are modulated significantly under these conditions. Addition of prostacyclin or sodium nitroprusside for 2 min before stimulation with thapsigargin results in the same dramatic reduction in thromboxane-€3, formation. Again, this situation is somehow reflected (Fig. 2 b) when studying cytosolic-calcium alterations under these conditions. Unstimulated assays contain approximately 1.5 ng thromboxane B,/ ml platelets. In the presence of BM13.177, prostacyclin or sodium nitroprusside, the thromboxane levels ranged only 2 3-fold above reference values, indicating that addition of a thromboxane-receptor blocker, or incubation with prostacyclin or sodium nitroprusside drastically reduced the ability of thapsigargin to cause thromboxane-B, formation. To characterize the effect of CAMP- or cGMP-elevating agents on intracellular calcium homeostasis in relation to the two-pool model, we investigated the redistribution of calcium already released (Fig. 3a and b). Using a platelet suspension in a nominally calcium-free solution with a blocked thromboxane receptor, (prior incubation with BM13.177), addition of thrombin produced a sharp immediate intracellularcalcium increase. After reaching a peak maximum, a slow decline in the Fura-2 signal was observed. Introducing thapsigargin (Fig. 3a, C) transiently releases stored calcium. Increasing the level of cyclic nucleotides (Fig. 3a, B), by adding prostacyclin or sodium nitroprusside after cytosolic calcium had been elevated by thrombin, accelerated the decline in the Fura-2 signal, with sodium nitroprusside being more effective compared to prostacyclin. Interestingly, by applying thapsigargin later, stored calcium was released again to the same extent (peak height) as in control incubations without altered cyclic-nucleotide levels. The same experimental setup described in Fig. 3a was used in Fig. 3b, with the exception of carrying out the experiments in the presence of a blocked thromboxane receptor in combination with all extracellular calcium being removed by addition of EGTA. After stimulation with thrombin, we observed a rapid intracellular-calcium increase followed by a relatively fast and nearly complete calcium uptake into intra-
cellular stores. Stored calcium can be released by subsequent addition of thapsigargin. In a situation excluding extracellular-calcium influx, the intracellular-calcium increase after thrombin shows the same peak height, but resequestration is much faster. Addition of prostacyclin or sodium nitroprusside (Fig. 3b, B) further accelerates calcium reuptake, again with the tendency of sodium nitroprusside to be more potent compared to prostacyclin. Applying thapsigargin (Fig. 3b, C) again released stored calcium, without any differences between individual assays. For comparison, we included the calcium trace produced by thapsigargin in the presence of BMl3.177 and EGTA when added alone, without a prior incubation employing thrombin as the first stimulus. Thapsigargin alone only releases about half of the calcium compared to thrombin. Cyclic nucleotides modulate thapsigargin-induced calcium influx
Calcium influx into platelets was measured using the manganese technique described by Sage and coworkers [31]. Stimulation of a receptor-operated calcium channel will result in stimulation of Mn2 influx associated with quenching of the dye Fura-2. Exposure of platelets to 1 pM thapsigargin augmented the rate of M n 2 + entry when compared to the rate measured during the control period before thapsigargin addition (Fig. 4a, trace 1). Blocking the thromboxane receptor with the receptor antagonist BM13.177 significantly reduces Mn2+ entry after adding thapsigargin (Fig. 4a, trace 2 versus 1). Inclusion of prostacyclin with BM13.177 abolishes the residual manganese entry (Fig. 4a, trace 3) after addition of thapsigargin. The same inhibitory activity of prostacyclin was demonstrated when BM13.177 was omitted (data not shown). In the same set of experiments we investigated the effect of sodium nitroprusside on thapsigargin-induced Mn2 influx, as shown in Fig. 4b. As indicated, thapsigargin-initiated M n 2 + influx is dramatically reduced in the presence of BM13.177 and is totally inhibited by further addition of sodium nitroprusside. When BM13.177 was omitted from the experiment, the same inhibitory activity of sodium nitroprusside was observed (data not shown). In this respect, there is no difference between either CAMP or cGMP. For comparison, we studied the thrombin-mediated Mn 2 + influx as presented in Fig. 5. Fig. 5a shsws that addition of thrombin and +
+
61 1
a
b 0 5mM EGTA lOOuM BM13177
100pM L
W U
z
W
10 U
3
U
2min
-A
ThrlOlUlmil
.___ A ThiIOlUlnll
A ThrlOlUlmll
E
-A
ThrIOlU/mil A Thr[OlU/mll ......__A:ThrlOlU/mII A TG[lpMI
I C TG[lpMI
E PG1211pg/mil E SNP1100pMl
____
C TG[lpMI C TGIlpMI
B. I C TGIlpMI B-PGl,Ilpg/mil C TGIlpM1 B .SNP[lOOpMI C TGIlpMI
Fig. 3. Cyclic nucleotides accelerate calcium redistribution into thapsigargin-sensitive calcium compartments. Fura-2-loaded platelets were challenged with various compounds at indicated concentrations according to the protocol given, in the presence of either BM13.177 (a) or BM13.177 plus EGTA (b). Intracellular calcium was measured as described under Experimental Procedures. TG, thapsigargin; PG12, prostacyclin; SNP, sodium nitroprusside; Thr, thrombin.
a
thrombin compared to thapsigargin, an inhibitor of the Ca2 ATPase, especially in a situation not allowing the self-amplification mechanism to be intact. Calculating the initial rate of Fura-2 quenching after using thapsigargin as an agonist (100% control value), addition of BM13.177 reduces this initial rate to values below 20% (80% inhibition). Investigating calcium entry and calculating the rate of [Ca2+Iiincrease during the first 1 0 s after Ca2+ addition revealed that the initial rate after addition of thapsigargin (100% control value) was only reduced by about 40% in the presence of BM13.111. These results indicate differences between [Ca2'Ii rise and Mn2+ entry. + -
0 0 DISCUSSION @ @ @)
I B TGIlpMI A BM13177IlOOpMI 8 - TGIlpMI .. ..-.- A BM13177[XX)pMI 8 T G I l p M I PG12 [lpg/rnll
-A
@ @ @)
I 6 TGflpMI A BM131771WOpMI 8 TGlIpMI -.-_...-. A EM13177[XX)pM1 B T G I l p M I SNP I100pMI
-A.
Fig. 4. BM13.177, prostacyclin and sodium nitroprusside inhibit Fura2 fluorescence in the presence of Mn2+. The fluorescence emission (excitation wavelength, 362 nm) of Fura-2-loaded platelets was recorded in the presence of 500 pM extracellular M n Z + ,as indicated under Experimental Procedures. At the various points indicated, different additions at the indicated concentrations were made. TG, thapsigargin; PG12, prostacyclin; SNP, sodium nitroprusside.
thapsigargin resulted in a similar slope of manganese influx. Influx was maximal and could not be further increased by subsequent addition of the other corresponding agonist, thrombin or thapsigargin. As indicated in Fig. 5b, addition of the thromboxane-receptor antagonist BMI 3.177 significantly inhibits thapsigargin-induced manganese influx, whereas thrombin-initiated manganese influx is not affected (Fig. 5b, compare traces 1 and 2). The slow Mn2+ influx caused by thapsigargin in the presence of BM13.177 is dramatically enhanced by the addition of a receptor agonist like thrombin (Fig. 5b, C). Therefore, manganese-influx experiments also reveal differences in the action of a receptor agonist like
Calcium release from intracellular stores and calcium influx Previous findings revealed the structural separation of a receptor-sensitive calcium-release compartment versus a compartment showing a sensitivity towards a Ca2+-ATPaseinhibitor like thapsigargin [I I]. Although earlier data suggested that thapsigargin empties the same calcium pool released by inositol 1,4,5-trisphosphate 1321, it became evident that thapsigargin also affects inositol-l,4,5-trisphosphate-insensitive compartments [ l l , 18-20, 33, 341. In platelets with an intact self-amplification system, thapsigargin. by acting not only as a Ca2+-ATPaseinhibitor but also indirectly as a receptor agonist, produces calcium traces similar to thrombin. This suggests that a receptor agonist and a Ca2+-ATPase inhibitor are indistinguishable in releasing calcium from intracellular stores. With a blocked thromboxane receptor, the agonistinduced calcium increases were always larger than those evoked by thapsigargin (Fig. 1). However, the magnitude of thapsigargin-induced Ca2+ increases varies considerably among different cell types, and in platelets it is also influenced by the endogenous formation of thromboxane A2. The time course of the progressive decay of the maximal thrombininduced Ca2 transient after thapsigargin treatment indicates that the agonist-sensitive compartment also shows a leakage +
61 2
b
a
@ @
B TG[lpMI I B Thr[OlU/mll I
-A
C Thr[OlU/rnll C TGIlpMI
B TGIlpMI @ -A:BM13177[1WpMi @ ----A BM13177[100pMI B ThrlOlU/rnll
C-ThrIOlUlrnll C TGllpMI
Fig. 5. Thapsigargin- and thrombin-induced manganese influx and its modulation by BM13.177. Mn2+ influx in platelets was measured as described under Experimental Procedures and explained in Fig. 4. At the indicated points, either BMI 3.1 77, thapsigargin (TG) or thrombin (Thr) were added at the concentrations indicated.
which is considerably slower than the leakage of the Table 1. Thrornboxane-BZ formation in washed human platelets. thapsigdrgin-sensitive compartment. This effect can not be Washed human platelets (5 -8 x lo8 platelets/ml) were taken during explained simply by removal of Ca2+ from the cells because original Furd-2 recordings. Samples were processed according to ExProcedures and thromboxane BZwas determined accordthe thrombin signal is more stable. It might indicate that the perimental ingly. Results are expressed as percentage of a control value in each thapsigdrgin-sensitive compartment may serve as a reservoir experiment, set as loo%, expressed as percentage of thromboxane-B2 for refilling the agonist-sensitive part. In this respect, the re- (TxB,) formation compared to the response towards thapsigargin. sults fit a proposal of Menniti and colleagues [35] describing All samples used 1 pM thapsigargin. Means 5 SD from four experthe inositol-I ,4,5-trisphosphate-sensitivestore as one com- iments are given. Samples were taken according to the experimental partment involved in Ca2' uptake and another compartment design presented in the corresponding figure, always 4 min after the mainly involved with Ca2 release. Our results demonstrate addition of thapsigargin. Prostacyclin (PG12) sodium nitroprusside that despite a prolonged incubation with thapsigargin in the (SNP) and BM13.177 were incubated for 2 min. presence of extracellular EGTA, the structural separation of Sample TxB~ two calcium-release compartments remains intact. formation Studying the functional relationship between Ca2' entry and intracellular-Ca2+pools in human platelets, thapsigargin % and thrombin produce equivalent Mn2+influx (Fig. 5 ) as long 100 as the self-amplification system remains intact. Preventing the Control 100 pM BM13.177 1.3 k 0.52 formation of thromboxane A2 reduces thapsigargin-induced 1 pg/ml prostacyclin 0.6 k 0.47 M n2+influx compared to that stimulated by thrombin. Intra- 100 pM sodium nitroprusside 2.1 0.50 cellular-calcium mobilization [I I], as well as the Ca2+-influx studies, indicate that thapsigdrgin alone, without endogenous formation of thromboxane A,, releases the part of the intracellular calcium associated with only a marginal M n 2 +influx, Cyclic nucleotides modulate thapsigargin-inducedcalcium although showing only partial inhibition of C a 2 +entry (Ca2' release overshoot). Our results revealed that thapsigargin-sensitive As previously reported and also shown here, thapsigargin intracellular-calcium compartments only play a minor role in opening channels allowing manganese influx (Fig. 5), com- effectively releases calcium and causes thromboxane-B, forpared to a classical receptor agonist like thrombin. Therefore, mation (Fig. 2; Table 1) by the ability of platelets to amplify in platelets, not only capacitative Ca2+ entry [37] originating the signal, resulting in the formation of a feed-forward actifrom empty thapsigargin-sensitive compartments, but also an vation signal. Cutting off the self-amplification system by additional component mediated by receptor activation and blocking the thromboxane receptor inhibits thapsigargin-inemptying of the related store seems to be involved in allowing duced thromboxane-B2 formation and modifies calcium reMn2+ influx. Similar observations, that simply emptying the lease significantly. Applying agents like prostacyclin or soinositol-l,4,5-trisphosphate-sensitive Ca2 stores did not lead dium nitroprusside, known to generate cAMP or cGMP inside to a significant stimulation of Ca2 entry, have already been the cell by activation of adenylate cyclase or soluble guanylate observed using hepatocytes [37, 381. Taking into account the cyclase, respectively, caused a similar modification of Ca2+-overshootmodel compared to M n 2 + entry, one should thapsigargin-induced calcium signals due to BM13.177. consider the possibility that emptying thapsigargin-sensitive Measuring thromboxane-Bz formation indicated that cAMP stores may open a specific Ca2' channel, not permeable to and cGMP inhibited the endogenous formation of this lipid Mn2' [39]. This would suggest the existence of two Ca2'- mediator and thereby blocked the self-amplification mechanentry pathways, with only one allowing the influx of M n 2 + . ism. Inhibition of phospholipase-C-mediated cell activation +
+
+
61 3 has been described for both cAMP and cGMP [40,41]. Therefore, prostacyclin and sodium nitroprusside mainly modulate the action of thapsigargin originating from the secondary ability of the Ca2+-ATPaseinhibitor when thapsigargin also acts as an indirect receptor agonist. Small differences between BM13.177 and prostacyclin/sodium-nitroprussidemay originate from calcium extrusion or indicate incomplete inhibition of the thromboxane receptor. Interestingly, thapsigargin still increased the intracellular calcium to an elevated plateau in the presence of the cyclic-nucleotide-forming compounds, indicating that this new steady state is not antagonized by plasma-membrane pump activation. Considering the twopool-compartment model, further experiments (Fig. 3) revealed that cyclic nucleotides stimulate calcium reuptake into the thapsigargin-sensitive intracellular store. This effect is more pronounced in the presence of extracellular calcium, with sodium nitroprusside being more effective than prostacyclin. Additional results (Fig. 4) support the concept that cAMP and especially cGMP are potent inhibitors of C a 2 + influx, as studied using human platelets [28, 421. This is also evident using thapsigargin to induce manganese influx as an indication of Ca2+ entry. The same results were observed employing thrombin (data not shown). Furthermore, the accelerated reversal of an elevated cytosolic-calcium level reported by MacIntyre et al. [27] was confirmed, indicating that uptake into the thapsigargin-sensitive compartment occurs, obviously not associated with calcium extrusion. Therefore, one can speculate that the thapsigargin-sensitive Ca2 ATPase is regulated by cAMP/cGMP taking part in the complex, an efficient way of cyclic nucleotides inhibiting complete platelet activation [43]. Our experiments reveal the interference of the cyclic nucleotides CAMP and cGMP with various calcium signals evoked by thapsigargin. Increased levels of cyclic nucleotides inhibit the self-amplification system of human platelets, not allowing the formation of thromboxane B2 after addition of thapsigargin. The Iack of a positive-feedback loop alters the calcium-release properties of thapsigargin compared to thrombin. By blocking the thromboxane receptor or adding prostacyclin or sodium nitroprusside, thapsigargin releases only part of the intracellular calcium compared to a normal receptor agonist like thrombin. Manganese influx, used as a means to monitor C a 2 + entry, is inhibited, and cyclic nucleotides accelerate the redistribution of cytosolic calcium into the thapsigargin-sensitive calcium compartment.
9. 10. 11 12 13 14 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.
+ -
The study was supported by the Deutsche Forschungsgemeinschaft (SFB 156/A4). The expert technical assistance of B. Diewald is gratefully acknowledged.
REFERENCES 1. Lapetina, E. G. (1986) in Phosphoinositides and receptor mechanisms (Putney, J. W., Jr, ed.) pp. 271 -286, Alan R. Liss, Inc., NY. 2. Lapetina, E. G. (1 983) L(fe Sci. 32, 2069 - 20823. 3. Berridge, M. J. & Irvine, R. F. (1984) Nature312, 315-321. 4. Nishizuka, Y. (1984) Nature 308, 693 - 698. 5. Flower, R. J. & Blackwell, G. J. (1979) Biochem. Pharmacol. 25, 285 - 291. 6. Hamberg, M., Svensson, J. & Samuelsson, B. (1975) Proc. Natl Acad. Sci. 72, 2994-2998. 7. Siess, W., Cuatrecasas, P. & Lapetina, E. G. (1983) J . Bid. Chem. 258,4683 -4686. 8. Rink, T. J. & Hallam, T. J. (1984) Trends Biochem. Sci. 376,215219.
25. 26. 27. 28 29. 30.
Rink, T. J. (1986) Agents Actions Suppl. 20, 141-169. Brune, B. & Ullrich, V. (1991) FEBS Lett. 284. 1-4. Brune, B. & Ullrich, V. (1991) J . Bid. Chem. 266,19 232-19 237 Thastrup, O., Dawson, A. P., Scharff, O., Foder, B., Cullen, P. J., Drobak, B. K., Bjerrrum, P. J., Christensen, S. B. & Hanley, M. R.(1989) Agents Actions 27, 17-23. Thastrupp, O., Foder, B. & Scharff, 0. (1987) Biochem. Biophys. Res. Commun. 142, 654- 660. Jackson, T. R., Patterson, S. L. & Thastrup, 0. (1988) Biochem. J . 253, 81 - 86. Moore, G. A., McConkey, D. J., Kass, G . E. N., O'Brien, P. J. & Orrenius, S. (1987) FEBS Leu. 224, 331 -336. Kass, G. E. N., Duddy, S., Moore, G . A. & Orrenius, S. (1989) J . Bid. Chem. 264, 15 192- 15 198. Oldershaw, K. A. &Taylor, C. W. (1990) FEBS Lett. 274, 21421 6. Koshiyma, H. & Tashjian, A. H., Jr (1991) Biochem. Biophys. Res. Commun. 177, 553 - 558. Menniti, F. S., Bird, G. St. J., Takemura, H., Thastrup, O., Potter, B. V. L. & Putney, J. W., Jr (1991) J . Biol. Chem. 266,13 64613 653. Takemura, H., Ohshika, H., Yokosawa, N., Oguma, K. & Thastrup, 0. (1991) Biochem. Biophys. Res. Commun. 180, 1518- 1526. Yamanishi, J., Kawahara, Y. & Fukuzuki, H. (1983) Thromh. Res. 32, 183-188. Knight, D. E. & Scrutton, M. C. (1984) Nature 309, 66-68. Bushfield, M., McNicoI, A. & MacIntyre, D. E. (1985) Biochem. J . 232,261 - 27 1. Pannochia, A. & Hardisty, R. M. (1985) Biochem. Biophys. Res. Commun. 127,339-345. Sage, S. 0. &Rink, T. J. (1985) FEBS Lett. 188, 135-140. Kawahara, Y., Yamanishi, J. & Fukuzaki, H. (1984) Thromh. Res. 33, 203 - 209. MacIntyre, D. E., Bushfield, M. & Shaw, A. M. (1985) FEBS Lett. 188, 383 - 388. Morgan, R. 0. & Newby, A. C. (1989) Biochem. J . 258, 447454. Brune, B. & Ullrich, V. (1991) Mol. Pharmacol. 39,671 -678. Grynkiewicz, G., Poenie, M. & Tsein, R. Y. (1985) J. Biol. Chem. 260, 3440 3450. Sage, S. O., Merritt, J. E., Hallam, T. J. & Rink, T. J. (1989) Biochem. J . 257,923 - 926. Jackson, T. R., Patterson, S. I., Thastrupp, 0. & Hanly, M. R. (1988) Biochem. J . 253,81-86. Thastrup, O., Cullen, P. J., Drobak, B. K., Hanley, M. R. & Dawson, A. P. (1990) Proc. Natl Acad. Sci. USA 87, 24662470. Ely. J. A., Ambroz, C., Baukal, A. J., Brogger Christensen, S., Balla, T. & Catt, K. J. (1991) J . Biol. Chem. 266, 1863518641. Menniti, F. S., Bird, G. St. J., Takemura, H., Thastrup, O., Potter, B. V. L. & Putney, J. W., Jr (1991) J . Biol. Chem. 266,13 64613 653. Putney, J. W. Jr (1990) Cell Calcium 11, 611 -624. Kass, G. E. N., Llopis, J., Chow, S. C., Duddy, S. K. & Orrenius, S. (1990) J . Bid. Chem. 265, 17 486- 17 492. Hansen, C. A., Yang, L. &Williamson, J. R. (1991) J . B i d . Chern. 266, 18 573 - 18 579. Neher, E. (1 992) Nature 335, 298 - 299. Billah, M. M., Lapetina, E. G. & Cuatrecasas, P. (1979) Biochem. Biophys. Res. Commun. 90,92 - 98. Nakashima, S., Tohmatsu, T., Hattori, H., Okano, Y. & Nozawa, Y. (1986) Biochem. Biophys. Res. Commun. 135,1099-1104. Simon, M. F. & Chap, H. (1989) J . Cardiovasc. Pharmacol. 14, S106 S110. Rink, T. J. & Sage, S. 0. (1990) Annu. Rev. Physiol. 52, 431 490. ~
31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.
~
43.
